Development of a scalable process for CI-1034, an endothelin antagonist

Article abstract of DOI:10.1021/op034104g

A concise, convergent multikilogram synthesis of CI-1034 (1), a potent endothelin receptor antagonist, is described. A 15-step preparation from commercially available o-vanillin and benzenesulfonyl chloride employs a remarkably robust Suzuki coupling between a boronic acid and an aromatic sulfonate ester as the key synthetic step. A scalable route capable of producing multikilogram quantities of CI-1034 with no chromatographic steps is described in this contribution. Improvements to the process included using a 4-fluorobenzenesulfonate ester as a suitable substitute for the triflate group in the Suzuki reaction and the use of MgCl₂ as a substitute for TiCl₄ in a Dieckmann condensation to provide the benzothiazine dioxide core.

Full text of DOI:10.1021/op034104g

Organic Process Research & Development 2004, 8, 201−212
Development of a Scalable Process for CI-1034, an Endothelin Antagonist
Thomas E. Jacks,*,† Daniel T. Belmont,†,‡ Christopher A. Briggs,† Nicole M. Horne,† Gerald D. Kanter,‡
Greg L. Karrick,†,§ James J. Krikke,#,¶ Richard J. McCabe,†,X Jason G. Mustakis,†,∇ Thomas N. Nanninga,†,§
Guy S. Risedorph,#,| Ronald E. Seamans,† Richard Skeean,‡ Derick D. Winkle,†, and Thomas M. Zennie#,|
Pfizer Global Research and DeVelopment, Holland Chemical Research and DeVelopment, Holland Analytical Research
and DeVelopment, 188 Howard AVenue, Holland, Michigan 49424, U.S.A., and Ann Arbor Chemical Research and
DeVelopment, 2800 Plymouth Road, Ann Arbor, Michigan 48105, U.S.A.
Abstract:
partner, an aryl moiety activated as a boronic acid 2a or
zincate 2b, for coupling via Suzuki or Negeshi protocol. The
other entity, a triflate 3 or other similarly activated structure,
can be seen as the “southern half” section of the molecule.
The northern half can be further disconnected to aldehyde
4, which is easily derived from commercially available
o-vanillin (5). The southern half activated ester can be
obtained from heterocycle 6 which can be synthesized in
three steps from commercially available benzenesulfonyl
chloride and 2-(trifluoromethyl)aniline.
A concise, convergent multikilogram synthesis of CI-1034 (1),
a potent endothelin receptor antagonist, is described. A 15-step
preparation from commercially available o-vanillin and ben-
zenesulfonyl chloride employs a remarkably robust Suzuki
coupling between a boronic acid and an aromatic sulfonate ester
as the key synthetic step. A scalable route capable of producing
multikilogram quantities of CI-1034 with no chromatographic
steps is described in this contribution. Improvements to the
process included using a 4-fluorobenzenesulfonate ester as a
suitable substitute for the triflate group in the Suzuki reaction
and the use of MgCl2 as a substitute for TiCl4 in a Dieckmann
condensation to provide the benzothiazine dioxide core.
Results and Discussion
Northern Half Strategy. The desired aryl boronic acid
2a was prepared from commercially available o-vanillin in
a seven-step synthetic sequence. Bromination of o-vanillin
(5) via a modification of a literature preparation3 provided
bromide 7 in 92% yield with only minor amounts of
dibromination and isomeric impurities (Scheme 2). Cleavage
of the aryl-methyl ether to the required bromocatechol 8,
however, presented a challenge due to the harsh conditions
usually required for this deprotection. Typical lab-scale
deprotection using BBr3 in CH2Cl2 was fast and efficient,
and we used this reagent in the pilot plant twice on fairly
large scale to move the project forward quickly. It was clear
from the start that this method was not a solution in the long
term due to the high cost of the reagent, the cryogenic
reaction conditions required, and concern with the energetics
of the quench on a manufacturing scale. Other methods
examined to perform this cleavage included AlCl3/pyridine,4
LiCl/DMF,5 and TMSCl/NaI6 and met with various levels
of success (Table 1). The method that proved most efficient
was the use of HBr in acetic acid at 90-100 °C.7 Although
this method required a large excess of HBr (typically 4 mol
equiv) under fairly harsh conditions, catechol 8 was isolated
in 81.9% yield after purification by recrystallization. Alky-
lation of the relatively unstable catechol 8 to aldehyde 4 was
conducted immediately and accomplished via treatment with
CH2Br2/K2CO3 in DMF.8 Although a number of other
alkylation methods were examined (CH2I2, K2CO3, DMF;9
Introduction
In support of our endothelin receptor antagonist program1
for the treatment of primary pulmonary hypertension and
congestive heart failure, we required multikilogram quantities
of CI-1034 (1), a novel ETA receptor antagonist. This
molecule, although achiral, possesses a complex structure
due to a highly functionalized aryl moiety. To meet the needs
of the project, a scalable route to CI-1034 was required to
quickly deliver kilograms of material in support of toxicology
research and clinical trials. As the discovery route2 appeared,
overall, to be a fairly workable synthetic approach, we used
this route as a solid starting point for the preparation of CI-
1034.
A retrosynthetic analysis of 1 (Scheme 1) reveals an
obvious disconnection to give a “northern half” coupling
* Corresponding author. Thomas E. Jacks, J-STAR Research, Inc. 3001
Hadley Rd., South Plainfield, NJ 07080. E-mail: tjacks@jstar-research.com.
Telephone: (908)-791-9100.
† Holland CRD.
# Holland ARD.
‡ Ann Arbor CRD.
§ Current address: Pfizer Global Research and Development, 7000 Portage
Rd., Kalamazoo, MI 49001.
| Current address: Lilly Global Process Chemical Research and Development,
Lilly Corporate Center, Indianapolis, IN 46285.
Current address: Pleotint LLC, 7705 W. Olive Rd., W. Olive, MI 49460.
¶ Current address: Department of Chemistry, 327 Padnos Hall, Grand Valley
State University, Allendale, MI, 49401.
X Current address: Pfizer Global Manufacturing, 685 3rd Ave., New York,
NY 10017.
(3) de Paulis, T.; Smith, H. E. Synth. Commun. 1991, 21, 1091.
(4) Lange, R. G. J. Org. Chem. 1962, 27, 2037.
(5) Bernard, A. M.; Ghiani, M. R.; Piras, P. P.; Rivoldini, A. Synthesis 1989,
287.
∇ Current address: Pfizer Global Research and Development, MS8118D 411
Eastern Point Rd, Groton, CT 06340.
(1) (a) Doherty, A. M. J. Med. Chem. 1992, 35, 1493. (b) Ray, A.; Hegde, L.
G.; Chugh, A.; Gupta, J. B. Drug DiscoVery Today 2000, 5, 455.
(2) Bunker, A. M.; Cheng X.; Doherty, A.; Lee, C.; Repine, J. T.; Skeean, R.;
Edmunds, J. J.; Kanter, G. D. U.S. Patents 6,265,399 and 6,252,070, 1998.
(6) Olah, G. A.; Narang, S. C.; Gupta, B. G. B.; Malhotra, R. J. Org. Chem.
1979, 44, 1247.
(7) Horton, W. J.; Spence, J. T. J. Am. Chem. Soc. 1958, 80, 2453.
(8) Danishefsky, S. Lee, J. Y. J. Am. Chem. Soc. 1989, 111, 4829.
10.1021/op034104g CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/02/2004
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
201

Scheme 1
Scheme 2 a
Table 2. Alkylation of catechol 8
method
scale
yield (%)
CH2I2/K2CO3/DMF
CH2ClBr/Cs2CO3/DMF
lab scalea
lab scalea
70
75
CH2Br2/K2CO3/DMF
222 kg
68.4
a Lab-scale yields reported, as scale-up yields were poor.
Scheme 3 a
a Reagents: a) Br2, KBr, HOAc, H2O, 92% b) BBr3, CH2Cl2, 85% c) HBr,
HOAc, 82% d)CH2ClBr, Cs2CO3, DMF, 75% e) CH2Br2, K2CO3, DMF, 68% f)
MeMgCl, 91%.
a Reagents: a) Et3SiH, CF3CO2H, b) distillation, 55% overall yield.
MeLi) it was discovered that the order of reagent introduction
was crucial to the success of this transformation. Thus,
addition of a solution of aldehyde 4 in THF to a cold solution
of MeMgCl in THF provided alcohol 9a in 91.1% isolated
yield with no detectable amount of 9b.
The reduction of alcohol 9a to furnish the desired bromide
10 proved to be the most challenging transformation in the
entire synthesis. The original route used a silane reduction
sequence (Et3SiH/CF3CO2H)13 and our first pilot campaign
employed this method. However, a full equivalent of the
triethylsilyl ether 11 is produced in this reaction, which
cannot be separated from the desired product by distillation
(Scheme 3). Because of the high cost of this approach, this
method was not considered suitable for the long term, and a
number of other methods were investigated (Scheme 4). A
one-step method using excess TMS-Cl/NaI in MeCN14 was
effective on a small scale, but not with multigram quantities.
Table 1. Cleavage conditions for aryl methyl ether 7
method
yield (%)
comment
BBr3/CH2Cl2
AlCl3/pyr./MeCN
LiCl/DMF
TMSCl/NaI/MeCN
HBr/HOAc
85a
86.9a
41.4b
traceb
81.9a
98.6% purity
10% SM remains
sig. decomposition
decomposition
98.4% purity
a Isolated yield. b HPLC yield.
and CH2ClBr, Cs2CO3, DMF10) the CH2Br2/K2CO3 in DMF
method consistently proved superior on scale in the pilot
plant (Table 2). The crude aldehyde was recrystallized from
heptane/ethyl acetate to provide 4 in a modest 68.4% isolated
yield.
With 4 in hand, the crucial one-carbon homologation/
reduction sequence was explored. Surprisingly, employing
a typical procedure where the MeMgCl solution (3 M in
THF) was added directly to a solution of 4 provided the
required alcohol 9a, but with 6-8% of the reduction product,
alcohol 9b!11 While looking into alternatives (TiMeCl3,12
(11) Wilk, B. K.; Helom, J. L.; Coughlin, C. W. Org. Process Res. DeV. 1998,
2, 407.
(12) Reetz, M. T. In Organometallics in Synthesis; Schlosser, M., E., Ed.; Wiley
and Sons: New York, 1994; p195.
(13) (a) West, C. T.; Donnelly, S. J.; Kooistra, D. A.; Doyle, M. P. J. Org.
Chem. 1973, 38, 2675. (b) Mayr, H.; Dogan, B. Tetrahedron Lett. 1997,
38, 1013.
(9) Tanaka, M.; Mukaiyama, C.; Mitsuhashi, H.; Wakamatsu, T. Tetrahedron
Lett. 1992, 33, 4165.
(10) Zelle, R. E.; McClellan, W. J. Tetrahedron Lett. 1991, 32, 2461.
(14) Sakai, T.; Miyata, K.; Utaka, M.; Takeda, A. Tetrahedron Lett. 1987, 28,
3817.
202
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Scheme 4 a
a Reagents and conditions: a) excess TMS-Cl, NaI, MeCN, 65%; b) TMS-Cl, NaI, MeCN; c) Zn, HOAc, 52% (for b and c); d) SOCl2, heptane, 93%; e) ZnCl2,
NaBH3CN, 76%; f) ZnCl2, NaBH4, 70% from 9a.
A more reliable and scalable approach was to prepare iodide
12 and reduce to 10 with Zn/HOAc.15 However, this
approach resulted in a significant amount of dimer 13. To
avoid this problem, the alcohol was instead converted to the
chloride 14 using a variety of known conditions (SOCl2,16
MsCl/Et3N,17 HCl (g),18 TMSCl/LiCl19). We settled on the
SOCl2 for cost and ease of processing reasons and reduced
with Zn/HOAc15 to provide excellent yields of the desired
bromide 10 without Wurtz-coupled products. Although this
method was fine on a laboratory scale, we were concerned
about using a Zn metal reduction on pilot-plant scale and
decided not to pursue this approach further.
We turned our attention to metal hydride reducing agents
to accomplish the desired reduction. A modification of a
literature method using NaCNBH3/ZnCl220 worked quite well
on a laboratory scale, and we were able to successfully scale
up this chemistry in the pilot plant, preparing bromide 10 in
70% yield (from alcohol 9a via chloride 14) after isolation
by distillation. We were, however, concerned about the
cyanide waste stream produced with this chemistry. Fortu-
nately, we discovered that the less expensive and cyanide-
free NaBH4 reducing agent in tandem with ZnCl2 gave
similar results. Using the ZnCl2/NaBH4 system in heptane/
THF we were able to produce bromide 10 in 70% yield in
two steps from alcohol 9a on 60-70-kg scale in the pilot
plant. The entire six-step sequence from o-vanillin provided
bromide 10 in a modest 32.7% overall yield.
assembly (see Scheme 5). Treating 5 with 2 equiv of
MeMgCl gave an alcohol as a complex mixture, that was
reduced with H2 and Pd/C and brominated with NBS/H2-
SO4 (cat.) in THF21 to provide crude 15 in 91% yield. Ether
cleavage to catechol 16 and alkylation completed this
sequence to the desired bromide 10 in an overall yield of
35-55% for the crude bromide 10. However, because the
yield in the last two steps was inconsistent, we were reluctant
to take this process to the pilot plant.
In a similar strategy, 2-ethylphenol 17 was used as a
starting material, establishing the ethyl group from the outset.
Bromination with either Br2 in heptane or NBS/H2SO4(cat.)
in THF21 provided the desired bromophenol 18. Formylation
with hexamethylenetetramine/(CH2O)x/H2SO4/HOAc22 or
MgCl2/(CH2O)x/Et3N23 provided the required aldehyde 19 in
modest yield (60-70%), depending on the exact conditions.
Dakin oxidation24 with H2O2/NaOH provided catechol 16
which could be alkylated as described above. However, once
again the yields varied wildly in the alkylation step regardless
of the method of preparation. This was likely due to the
instability of the electron-rich catechol 16 in base, where
significant evidence of decomposition was usually observed
and alkylation of the ring was likely. Also, all alternate routes
to 10 outlined in Scheme 5 provided intermediates that were
either oils or low-melting solids, making purification of
intermediates difficult. Because of the harsh formylation
conditions and the low or unpredictable yield in several of
the steps, these alternate routes offered no real cost advantage
and were dropped from further consideration. Although we
were not completely satisfied with the described route to
bromide 10, we had shown that this chemistry was feasible
on-scale and were confident that processing improvements
could be made over time.
Although the route to bromide 10 just described proved
to be scalable, several different approaches were examined
to prepare this compound. One approach used o-vanillin as
the starting material but employed a different sequence of
(15) Shue, Y.-K.; Carrera, G. M., Jr.; Nadzan, A. M. Tetrahedron Lett. 1987,
28, 3225.
(16) Page, G. A.; Tarbell, D. S. J. Am. Chem. Soc. 1953, 75, 2053.
(17) Lynch, J. E.; Choi, W.-B.; Churchill, H. R. O.; Volante, R. P.; Reamer, R.
A.; Ball, R. G. J. Org. Chem. 1997, 62, 9223.
(18) Boekelheide, V.; Vick, G. K. J. Am. Chem. Soc. 1956, 78, 653.
(19) Olah, G. A.; Gupta, B. G. B.; Malhotra, R.; Narang, S. C. J. Org. Chem.
1980, 45, 1638.
(21) Coleman, R. S.; Grant, E. B. J. Org. Chem. 1991, 56, 1357.
(22) Weerawarna, S. A.; Guha-Biswas, M.; Nelson, W. L. J. Heterocycl. Chem.
1991, 28, 1395.
(23) Hofslokken, N. U.; Skattebol, L. Acta Chem. Scand. 1999, 53, 258.
(24) Wriede, U.; Fernandez, M.; West, K. F.; Harcourt, D.; Moore, H. W. J.
Org. Chem. 1987, 52, 4485.
(20) Kim, S.; Kim, Y. J.; Ahn, K. H. Tetrahedron Lett. 1983, 24, 3369.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
203

Scheme 5 a
a Reagents: a) MeMgCl; b) H2, Pd/C, 96% for steps a and b; c) NBS, THF, H2SO4(cat.), 95%; d) AlCl3, pyr., NaI, MeCN, 84% e) CH2Br2, K2CO3, DMF, 76%
crude yield; f) C6H12N4, (CH2O)x, H2SO4, 60%; g) MgCl2, (CH2O)x, Et3N, MeCN, 70%; h) NaOH, H2O2, 90% (crude yield).
Scheme 6 a
Southern Half Strategy. Heterocycle 6 was envisioned
as a key intermediate in construction of the benzothiazine
dioxide core structure of CI-1034. Starting from readily
available benzenesulfonyl chloride and 2-(trifluoromethyl)-
aniline, sulfonamide 20 was prepared in 97% yield with 99%
purity on large scale (Scheme 7).
The sulfonamide was metalated with 2 mol equiv of
n-BuLi at 0 to -10 °C and the resulting dianion 21 trapped
by reaction with CO2(g). This key step required the slow
a Reagents and conditions: a) n-BuLi, -70 °C; b) B(OC3H7)3; c) HOAc(aq),
95% for a-c.
addition of the dianion reaction mixture as a slurry in THF
The final step in the northern half sequence was prepara-
to a separate vessel of THF where CO2 was constantly being
tion of a suitably activated species for coupling with the
introduced. This procedure was carried out by bubbling CO2
southern half moiety. In the original synthesis, bromide 10
was activated for coupling by preparing the zincate 2b and
using it directly in the Negishi coupling step. We chose to
examine the preparation of boronic acid 2a (Scheme 6).
Boronic acids are generally stable and can be prepared and
isolated (and purified, if necessary) or used in situ as a
solution.25 Thus, a THF solution of bromide 10 at -60 to
into the mixture below surface level via a sparging tube
throughout the dianion addition. Whereas quenching 21 by
direct introduction of CO2 into the dianion slurry resulted
in up to 15% of ketone 22, by using this procedure no
detectable levels of 22 were produced. The ensuing lithium
carboxylate was converted directly to the desired cyclic
sulfonamide 6 by treatment with acetic anhydride and
-70 °C was treated sequentially with n-BuLi and triisopropyl
catalytic methanesulfonic acid and isolated in an overall yield
borate to provide, after quenching with aqueous acid, the
of 89.6% for the two steps.
desired boronic acid, but in a modest 70-80% yield and
contaminated with a number of impurities. Although we
could purify the boronic acid via recrystallization, we
examined this key reaction to see if we could improve the
impurity profile. It was found that better results could be
realized if the bromide was lithiated using an inverse addition
method. Thus, to a cold (-60 to -70 °C) solution of n-BuLi
in THF/hexanes was slowly added bromide 10, neat. After
stirring for 30 min, the mixture was treated with triisopropyl
borate. Warming to 0 °C and quenching with an aqueous
solution of acetic acid provided boronic acid 2a in excellent
yield (>95%) and purity. The boronic acid could then be
isolated or, preferably, used directly in the Suzuki coupling
step. From o-vanillin, the required northern half boronic acid
2a was prepared in seven steps with an overall yield of
Because sulfonamide 6 is a saccharin derivative, we
explored its preparation via commercially available sodium
saccharin (23) (Scheme 8). Reaction of sodium saccharin
with HCl(aq) and removal of water provided salt 24
(containing approximately 17% NaCl).26 Activation with
thionyl chloride, amide formation with 2-(trifluoromethyl)-
aniline, and finally activation/cyclization provided the desired
sulfonamide 6 in 70-75% isolated yield. In a similar fashion,
mixed anhydride 2527 (commercially available or easily
prepared) could be treated with thionyl chloride and 2-(tri-
fluoromethyl)aniline to provide 6 in 73% yield from sodium
saccharin in >98% purity. Although these yields were
slightly lower than the benzenesulfonyl chloride approach,
this attractive route avoided n-BuLi and the problems
31.6%.
(26) Clarke, H. T.; Dreger, E. E. Organic Syntheses; John Wiley & Sons: New
York, 1932; Collect. Vol. I, p 14.
(25) Zhao, D.; Xu, F.; Chen, C.; Tillyer, R. D.; Grabowski, E. J.; Reider, P. J.;
(27) Clarke, H. T.; Dreger, E. E. Organic Syntheses; John Wiley & Sons: New
Black, C.; Ouimet, N.; Prasit, P. Tetrahedron 1999, 55, 6001.
York, 1932; Collect. Vol. I, p 495.
204
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Scheme 7 a
a Reagents and conditions: a) pyridine, MeCN, then HCl(aq), 96.8%; b) i n-BuLi, ii CO2(g); c) Ac2O, CH3SO3H, 89.6% for b and c; d) CO2(g) added directly
to 21.
Scheme 8 a
MeCN could be performed. Reaction with methyl bromoac-
etate in MeCN with Na2CO3 as base provided diester 26 in
89% yield with slightly lower product purity (97%). Thus,
by tolerating slightly lower product purity a substantial gain
in ease of processing was realized (Scheme 9).
On the basis of this encouraging result, we briefly
explored a more direct preparation of monoester 27. Reaction
of sulfonamide 20 with n-BuLi, trapping with CO2, and
Fischer esterification in MeOH gave monoester 27 but in
modest yield. Using a different approach, treatment of
sulfonamide 20 with 2 equiv of n-BuLi and trapping with
methyl chloroformate provided direct access to the desired
monoester, but with 15-20% starting material, presumably
due to HCl in the methyl chloroformate. Both approaches
provided straightforward access to diester 26, and we felt
a Reagents and conditions: a) HCl(aq), b) i SOCl2, ii C7H6F3N, iii SOCl2,
70-75% for a-b; c) i distill, ii SOCl2; d) i SOCl2, ii C7H6F3N, 70-75%.
both approaches were viable and worthy of further investiga-
tion. Time restraints, however, prevented further develop-
ment.
associated with dianion 21. However, unpredictable gas
With 26 in hand, we proceeded to the key Dieckmann
evolution plagued this procedure regardless of the conditions
condensation28 that would provide the benzothiazine dioxide
examined. Due to the large scale that would be required and
core of CI-1034 (Scheme 10). The original route employed
the severe time constraints of the project, these routes to 6
were dropped from further development.
Preparation of diester 26 was accomplished in a straight-
forward manner by treatment of 6 with methyl bromoacetate
a Lewis acid-catalyzed approach to ketoester 28. Using 2.2
mol equiv of TiCl4 in CH2Cl2 with Et3N as base at -55 °C,
ketoester 28 (shown as the enol form) was cleanly and
efficiently prepared in 98% yield.29 However, the use of less
in a solution of 25% sodium methoxide in methanol. The
base, elevated temperatures, or different solvents resulted in
product precipitated from the reaction mixture upon cooling
drastically reduced yield and purity. Typically, we did not
in 80-85% yield and >99% purity. This chemistry was
isolate 28 but used it directly in the next step as a solution
performed a number of times in our pilot plant on a large
in CH2Cl2 after an aqueous work up and drying over MgSO4.
scale. However, this reaction resisted going to completion,
The yield of the TiCl4-catalyzed Dieckmann condensation
even upon addition of additional reagents or in different
was superb, but there are a number of drawbacks to
employing this reagent on a large scale. Although not
solvent systems. Monoester 27 is observed in the course of
the reaction and is easily isolated, and thus we explored a
expensive, TiCl4 requires special handling, low-temperature
two-step procedure to 26. Reaction of sulfonamide 6 with
conditions in CH2Cl2, and produced significant waste that
NaOMe in MeOH and quenching with HCl(aq) provided
we had to treat before disposal.
monoester 27 in 91% isolated yield. Treatment of the
Unfortunately, a Lewis acid-catalyzed approach to ke-
toester 28 seemed a certainty. When treated with strong bases
such as LDA, HMDSLi, NaH, etc., diester 26 undergoes a
monoester with methyl bromoacetate in MeCN with Na2-
CO3 as base and catalytic sodium iodide provided diester
26 in 90% yield (for the two steps) and >98% purity.
Alternatively, instead of employing two isolations, after
(28) Tanabe, Y. Bull. Chem. Soc. Jpn. 1989, 62, 1917.
reaction to form monoester 27 a solvent exchange with
(29) Patek, M. Collect. Czech. Chem. Commun. 1990, 55, 1223.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
205

Scheme 9 a
a Reagents and conditions: a) NaOMe, isolate or solvent exchange with MeCN, 91%; b) C3H5O2Br, Na2CO3, NaI, MeCN, 99%; c) i n-BuLi, ii CO2,H+ iii MeOH,
H+, 55% d) i n-BuLi, ii ClCO2Me, 75%.
Scheme 10 a
a Reagents and conditions: a) TiCl4, Et3N, CH2Cl2, -55 °C, 98.5%; b) MgCl2, DBU, MeCN, 89.7%; c) Tf2O, pyridine, CH2Cl2, 93.2%.
Table 3. Dieckmann condensation to prepare ketoester 28
using MgCl2 as Lewis acid catalyst with various bases and
solvents
rapid decomposition, even at very low temperatures, with
only minor amounts of 28 produced (Scheme 10). The use
of methoxide in selected solvents (such as toluene) worked
base
yield
(%)
product purity
(HPLC area %)
to some degree, but the yields were not attractive and this
approach was deemed unacceptable. The use of inexpensive
and easy-to-handle MgCl2 as a Lewis acid catalyst in
carbon-carbon bond forming reactions has been reported.30
Much to our delight, this salt, in conjunction with many
standard organic bases (using 3 mol equiv), was very
effective and provided an attractive alternative to TiCl4
(Table 3). Using MgCl2 and DBU in MeCN ketoester 28
was obtained in 89.7% isolated yield with excellent purity.
Although the yield and purity of ketoester 28 was slightly
lower, MgCl2 was demonstrated as a viable alternative that
eliminated the low-temperature chemistry, CH2Cl2 solvent,
and special handling and waste stream problems associated
with TiCl4.
solvent
MeCN
base
(mol equiv)
pyridine
5.0
NR
96a
starting material
recovered
98
MeCN
MeCN
MeCN
THF
THF
DME
DIPEA
DBU
Et3N
Et3N
DBU
Et3N
3.6
3.0
3.0
3.0
3.0
3.5
3.0
89.7a 98.5
99a
87a
83b
87b
44b
95
98
incomplete rxn
incomplete rxn
isolation issues
toluene DBU
a Isolated yield. b HPLC (area %) yield.
yield (for the two steps) with >99% purity. Overall, the yield
of triflate 3 was a very respectable 68% starting from
benzenesulfonyl chloride.
Preparation of triflate 3 from a CH2Cl2 solution of
ketoester 28 was straightforward,31 providing 3 in 93.2%
(30) Tamai, S.; Ushirogochi, H.; Sano, S.; Nagao, Y. Chem. Lett. 1995, 295.
(31) Stang, P. J.; Hanack, M.; Subramanian, L. R. Synthesis 1982, 85.
206
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Scheme 11 a
a Reagents and conditions: a) R-Cl, Et3N, MeCN 75-85%; b) 2a, Na2CO3, PdCl2(PPh3)2, PPh3, 85-95% (for 3, 29e, 29f).
Table 4. Catalyst, ligands, and bases for preparation of 30
equivalents of trithiocyanuric acid38 was quite effective,
reducing levels of palladium to <10 ppm. Thus, we were
not concerned with using palladium metal in the next-to-
last step of the synthetic sequence.
from 3 and 2a
catalyst
load
yield
catalyst
ligand
2 PPh3
(%)
base
(%) comment
Triflate 3 proved to be a superb coupling partner in the
assembly of CI-1034. However, triflic anhydride is very
expensive and has an “atom efficiency”39 of zero. Thus, we
felt it was imperative to find a suitable replacement if the
current synthetic strategy was to remain viable as a com-
mercial route. Significant advances have been realized
recently in Suzuki-type couplings,40 suggesting a number of
coupling strategies worthy of investigation. However, a
number of criteria had to be met: (1) the catalyst load must
not substantially increase due to the high cost of metal
catalysts, (2) the ligand must be commercially available and
non-proprietary, (3) the process with the new catalyst must
be robust on scale, and (4) the triflate substitute must be
inexpensive and easy to prepare.
PdCl2(PPh3)2
0.3
Na2CO3 94.6a best pilot
plant yield
Pd(OAc)2
4 PPh3
none
0.3
2.5
K2CO3 89a
PdCl2dppp33
Na2CO3 16b triflate
hydrolysis
NiI2dppp34
2 PPh3
2 PCy3
5
3
0.5
Na2CO3 5b
35
c
Pd(OAc)2
KF
K2CO3 27b triflate
7b
36
PdCl2P(oTol3)2 2 P(oTol)3
hydrolysis
Pd2(dba)3
PdCl2
2 PPh3
4 PPh3
none
0.8
0.5
Na2CO3 92a
K2CO3 60b
Pd/C37
K2CO3 55b triflate
hydrolysis
a Isolated yield. b HPLC yield. c Tricyclohexylphosphine.
A number of suitable southern half coupling entities (29
a-f) were prepared and evaluated in reactions with boronic
acid 2a. The coupling partners studied included the mesy-
late,41 tosylate,42,43 phosphate,43 benzenesulfonate,44 4-fluo-
robenzenesulfonate,43 and 4-chlorobenzenesulfonate esters.
Couplings with 2a were conducted using a variety of catalyst/
ligand combinations as guided by literature precedent (Table
5). The mesylate and phosphate groups were poor coupling
partners, primarily due to their instability in the basic aqueous
reaction mixture. The tosylate had limited success, but the
reaction was not fast enough to avoid precipitation of the
Coupling Strategy. With both the northern half boronic
acid 2a and the southern half triflate 3 in hand, we were
ready to assemble the two halves using a palladium-catalyzed
Suzuki coupling reaction. A solution of boronic acid 2a was
transferred into a vessel containing triflate 3, Na2CO3, and
the catalyst system of PdCl2(PPh3)2/2 PPh3. This catalyst
system was efficient and robust, with the reaction going to
completion in several hours with a mere 0.3 mol % catalyst
load. After an aqueous workup, ester 30 crystallized directly
from the complex solvent system in 90-95% isolated yield
(Scheme 11).
(33) Badone, D.; Cecchi, R.; Guzzi, U. J. Org. Chem. 1992, 57, 6321.
(34) Huffman, M. A.; Yasuda, N. Synlett 1999, 4, 471.
A number of other catalyst systems were reviewed for
this transformation (see Table 4), but none proved as reliable
and robust as the PdCl2(PPh3)2/2 PPh3 system. For example,
(35) Littke, A. F.; Dai, C.; Fu, G. C. J. Am. Chem. Soc. 2000, 122, 4020.
(36) Herrmann, W. A.; Brossmer, C.; O¨ fele, K.; Reisinger, C.-P.; Priermeier,
T.; Beller, M.; Fischer, H. Angew. Chem., Int. Ed. Engl. 1995, 34, 1844.
(37) Gala, D.; Stamford, A.; Jenkins, J.; Kugelman, M. Org. Process Res. DeV.
1997, 1, 163.
(38) Rosso, V. W.; Lust, D. A.; Bernot, P. J.; Grosso, J. A.; Modi, S. P.;
Rusowicz, A.; Sedergran, T. C.; Simpson, J. H.; Srivastava, S. K.; Humora,
M. J.; Anderson, N. G. Org. Process Res. DeV. 1997, 1, 311.
(39) Trost, B. M. Science 1991, 254, 1471.
(40) Kotha, S.; Lahiri, K.; Kashinath, D. Tetrahedron 2002, 58, 9633.
(41) (a) Zim, D.; Lando, V. R.; Dupont, J.; Monteiro, A. L. Org. Lett. 2001, 3,
3049. (b) Percec, V.; Bae, J.-Y.; Hill, D. H. J. Org. Chem. 1995, 60, 1060.
(42) Kubota, Y.; Nakada, S.; Sugi, Y. Synlett 1998, 183.
(43) (a) Badone, D.; Cecchi, R.; Guzzi, U. J. Org. Chem. 1992, 57, 6321. (b)
Sasaki, M.; Fuwa, H.; Ishikawa, M.; Tachibana, K. Org. Lett. 1999, 1, 1075.
(44) Zim, D.; Lando, V. R.; Dupont, J.; Monteiro, A. L. Org. Lett. 2001, 3,
3049.
32
Pd(OAc)2/4 PPh3 performed well, but was not as robust
(the reaction frequently would stop due to formation of
“palladium black”), as the PdCl2(PPh3)2/PPh3 system.
We were concerned that the use of palladium metal so
late in the synthesis would present a serious problem with
its removal from the final product. Recrystallization alone
was not effective nor was treatment with various commercial
carbons. However, we found that reaction with several molar
(32) Smith, G. B.; Dezeny, G. C.; Hughes, D. L.; King, A. O.; Verhoeven, T. R.
J. Org. Chem. 1994, 59, 8151.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
207

Scheme 12 a
a Reagents and conditions: a) i n-BuLi, ii B(OiPr)3, iii HOAc(aq) b) PdCl2(PPh3)2, 2 PPh3, Na2CO3, 87%.
Table 5. Results of leaving group, catalyst, and ligand for
coupling of 29a-f with 2a
scale-up of the Suzuki coupling was performed using PdCl2-
(PPh3)2/2 PPh3 as the catalyst system, providing ester 30 in
86.9% yield on a 50-g scale. This yield was comparable to
the yield obtained using triflate 3, but the reaction time was
significantly longer. Unfortunately, the project was canceled
before we had the opportunity to test this route to ester 30
in the pilot plant. Nonetheless, a viable substitute to the
triflate group was discovered and was demonstrated suc-
cessfully on a moderate laboratory scale (Scheme 12).
The final transformation in the synthesis was the saponi-
fication of ester 30 and isolation as the potassium salt, CI-
1034 (1). The original route employed a classic hydrolysis
in water/methanol followed by acidification and then titration
with stoichiometric KOH to obtain CI-1034. This simple
hydrolysis requires surprisingly forceful conditions, several
hours at reflux with a large excess of hydroxide. We desired
a more streamlined approach. A report in the literature
described the use of potassium trimethylsilanolate45 to
convert esters directly to potassium carboxylates. We found
that this reagent converted ester 30 directly to the potassium
salt in THF over several hours at room temperature. While
this was a surprisingly facile transformation, for unknown
reasons the CI-1034 produced was unsuitable for filtration
on a multikilogram scale. A solvent exchange with IPA was
required to obtain a product with acceptable physical
properties. Because of the high cost of this reagent and the
poor filterability of the ensuing salt, a more direct method
was investigated. On the basis of the success in a similar
molecule, a stoichiometric quantity of KOH in IPA with 10%
water was investigated. This method worked quite well as
did 2-butanol with 10% water, providing CI-1034 in 78-
83% yield after azeotropically removing most of the water.
This procedure allowed the efficient preparation and isolation
of CI-1034 with excellent purity and in an acceptable yield
(Scheme 13).
catalyst load yield
entry compound
X
catalyst
ligand
(mol %)
(%)
1
2
3
4
5
6
7
8
29a
29b
29b
29c
29d
29d
29e
29f
P(O)(OPh)2 PdCl2(dppf)
3.0
5.0
3.0
1.0
0.3
0.3
0.3
0.2
2a
2a
Ms
Ms
Ts
PdCl2(PPh3)2 2 PPh3
Pd(OAc)2 2 PCy3
PdCl2(PPh3)2 2 PPh3
15a
46a
46a
43a
87b
87b
4-ClPhSO2 PdCl2(PPh3)2 2 PPh3
4-ClPhSO2 Pd(OAc)2 2 PCy3
SO2Ph PdCl2(PPh3)2 2 PPh3
4-FPhSO2 PdCl2(PPh3)2 2 PPh3
a HPLC yield. b Isolated yield.
Table 6. Results of leaving group, catalyst, and ligand for
coupling of 2a with 29e and 29f to prepare 30
catalyst load isolated
(mol %) yield (%)
entry compound
catalyst
ligand
1
2
3
4
5
6
7
8
9
11
29e
29e
29e
29e
29e
29e
29f
29f
29f
29f
PdCl2(PPh3)2
PdCl2(PPh3)2 2 PPh3
Pd2(dba)3
Pd(OAc)2
Pd(OAc)2
PdCl2(dppf)
-
0.3
0.3
1.0
0.5
0.5
0.5
0.23
0.5
0.3
1.0
51b
87a
91a
72a
73a
31b
87a
2 PCy3
2 P(nBu)3
2 PCy3
-
PdCl2(PPh3)2 2 PPh3
PdCl2(dppf)
Pd(OAc)2
Pd2(dba)3
-
2 PCy3
2 PCy3
43b
91a
87a
a Isolated yield. b HPLC yield.
catalyst as palladium black. The 4-chlorobenzenesulfonate
coupled fairly effectively, but a major side reaction was
coupling at the 4-chloro position of the benzenesulfonate!
By far the two most promising entities proved to be
benzenesulfonate ester 29e and 4-fluorobenzenesulfonate
ester 29f. A series of parallel experiments were set up to
guide the research towards development of a synthetic
process (Table 6). Although benzenesulfonate 29e was
slightly less expensive to prepare, in general it required
longer reaction times and a larger catalyst load to drive the
reaction to completion. A key concern with this chemistry
was the precipitation of the catalyst as “palladium black”,
which stops the progress of the reaction. This problem
occasionally plagued both groups, but overall, 4-fluoroben-
zenesulfonate 29f provided more reliable and consistent
yields of 30 in coupling with boronic acid 2a. Several other
catalysts, such as Pd2(dba)3/2 PCy3 and Pd(OAc)2/2 PCy3
had better or comparable yields in a single experiment but
were unreliable for reasons that were not clear. A multigram
Conclusions
An efficient synthesis of CI-1034 was developed and run
in the pilot plant on a multikilogram scale. The process
involved eight linear steps (15 total synthetic steps) with an
overall yield of 50.4% starting from commercially available
materials and with no chromatographic separations. A total
of 200 kg of CI-1034 was prepared under cGMP conditions
in our pilot plant in support of this project. The key step, a
remarkably robust Suzuki coupling of an aryl boronic acid
and a triflate, was demonstrated on an 80-kg scale with a
(45) Laganis, E. D.; Chenard, B. L. Tetrahedron Lett. 1984, 25, 5831.
208
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Scheme 13 a
h at <10 psi maximum pressure. Upon reaction completion
(<2% starting material present) the mixture was cooled to
about 50 °C and the solvent reduced under vacuum by
approximately 50%. The mixture was cooled to ambient
temperature, diluted with water (700 L), and cooled to
0-5 °C for 2 h. The product was collected by centrifugation,
washed with water (2 × 100 L, each load), and recrystallized
(water wet) from heptane/ethyl acetate to yield 187.2 kg of
8. The relatively unstable and solvent-wet 8 (172.9 kg
calculated dry weight, 797 mol, 81.9%) was carried on to
the next step without drying or further characterization.
HPLC: 98.4%; MS m/z 217/219 (M + H)+.
a Reagents and conditions: a) KOTMS, THF, then IPA, rt, 60%; b) KOH,
2-butanol, water, 80 °C, 78.3%.
very low catalyst load of 0.3 mol %. Several key improve-
ments were made to decrease the cost and improve process-
ing. An efficient MgCl2-catalyzed Dieckmann condensation
was developed to replace TiCl4 and was demonstrated on
large laboratory scale. Suitable replacements for the expen-
sive triflate group, 4-fluorobenzenesulfonate ester, were
demonstrated on a modest laboratory scale in excellent yield.
Aldehyde 4. The solvent-wet catechol 8 (172.9 kg,
calculated dry wet), K2CO3 (216 kg, 1.56 kmol), DMF (865
kg, 916 L), and CH2Br2 (211 kg, 1.21 kmol) were charged
to a 2000 L reactor and heated to 85 ( 5 °C for 5 h. The
mixture was cooled to about 50 °C and vacuum distilled to
remove about half of the solvent. Water (500 L) was added,
the mixture was cooled to ambient temperature, and HOAc
(75 kg, 1.25 kmol) was slowly added followed by additional
water (500 L). The contents were cooled to 0-5 °C and held
at that temperature overnight (14 h). The crude product was
collected by centrifugation and washed with water; the
solvent-wet 4 was recrystallized from heptane/ethyl acetate
to yield, after vacuum-drying, 124.8 kg of 4 (68.4%) as a
tan-colored solid; mp: 107 °C; HPLC: 99.2% (by area %);
1H NMR (400 MHz, DMSO-d6) δ 9.96 (s, 1H), 7.44 (d, J )
2.0 Hz, 1H), 7. 39 (d, J ) 2.0 Hz,1H), 6.26 (s, 2H); 13C
NMR (100 MHz, DMSO-d6) 187,9, 150.9, 149.3, 123.5,
120.1, 116.7, 113.2, 104.5; MS m/z: 229/231 (M + H)+.
Alcohol 9a. To a 2000 L reactor were charged methyl-
magnesium chloride solution (189 kg of 3.0 M in THF; 561
mol) and THF (350 L), and the solution cooled to 0 ( 5 °C.
In a separate 2000 L vessel were charged aldehyde 4 (100.0
kg, 437 mol) and THF (500 L). The solution of aldehyde 4
was cooled to 0-5 °C and transferred to the reactor
containing the Grignard reagent over about 75 min, keeping
the temperature below 15 °C during the addition. Upon
reaction completion, the reaction was quenched by careful
addition of cold, aqueous HCl (2.3 M, 560 L). The phases
were separated, and the organic phase was reduced under
vacuum and diluted with ethyl acetate (388 L). After stirring
with carbon/filtercel (4.0 kg/3.0 kg) and filtering, the solution
was reduced by vacuum distillation to a total volume of about
200-250 L. The solution was diluted with heptane (220 L)
and cooled to 0 ( 5 °C for 2 h. The product was isolated by
centrifugation, washed with heptane (2 × 55 L) and dried
under vacuum to yield 97.6 kg, (91.1%) of 9a as an off-
white solid. This material was used in the next step without
further purification. A sample for characterization was
recrystallized from heptane/ethyl acetate; mp: 74-75 °C;
HPLC: 96.1% (area %); 1H NMR (400 MHz, DMSO-d6) δ
7.07 (d, J ) 1.3 Hz, 1H), 7.03 (d, J ) 1.3 Hz, 1H), 6.05 (d,
J ) 13.6 Hz, 1H), 6.05 (d, J ) 13.6 Hz, 1H) 5.33 (d, J )
4.6 Hz, 1H), 4.78 (m, 1H), 1.31 (d, J ) 6.4 Hz, 3H); 13C
NMR (100 MHz, DMSO-d6) δ 148.0, 142.8, 130.2, 121.2,
112.1, 110.0, 62.8, 23.8; MS m/z: 244/246 (M+), 227/229
(M - OH)+.
Experimental Section
General Methods. Commercially available solvents and
reagents were used as received without further purification.
All reactions were performed under nitrogen unless otherwise
noted. 1H and 13C spectra were determined in DMSO-d6 using
a Varian INOVA 400 spectrometer at 400 and 100 MHz,
respectively. HPLC analysis performed using a Waters
Symmetry Shield RP8 column, 150 mm × 3.9 mm, 5 µm
using a water/acetonitrile gradient method at 2.0 mL/min at
210 nm. Mass spectrometry was performed using a Finnegan
4500 in DCI (isobutane) mode. Melting points were obtained
using a Thomas-Hoover capillary apparatus and were uncor-
rected. Parallel reactions were performed using either a
Radley’s Carousel Reaction Station or a Bohdan Process
Development Workstation.
Bromide 7. A 2000 L vessel was charged with o-vanillin
(160.0 kg, 1.05 kmol), acetic acid (800 kg), potassium
bromide (250 kg, 2.10 kmol), and water (440 L). The reaction
mixture was heated to 100 ( 5 °C, and bromine (175.0 kg,
1.09 kmol) was charged over the course of 55 min. After
stirring at 100 ( 5 °C for 90 min, the mixture was cooled
to about 70 °C and distilled under vacuum until the volume
was reduced to about 450 L total volume. Water (1000 L)
was charged and the mixture cooled to 0-5 °C. After 2 h at
this temperature the product was isolated by centrifugation,
washed with water (2 × 100 L), and dried for 21 h in a
vacuum oven at 50-55 °C. Yield: 222.7 kg (91.8%), mp:
122-124 °C, literature:5 128 °C; HPLC purity: 97.4%
(area %).
Bromocatechol 8. A 2000 L reaction vessel was charged
with bromide 7 (224.8 kg, 973 mol), acetic acid (1428.7 kg),
and water (100 L) and heated to 55 ( 5 °C. Under a vacuum
blank, anhydrous hydrogen bromide (210 kg, 2.60 kmol) was
charged over a period of 3 h. Upon completion of HBr
addition, the reaction mixture was heated at 85 ( 5 °C for
18 h, keeping pressure <10 psi by occasional venting. The
reaction mixture was cooled to 55 ( 5 °C, and an additional
charge of anhydrous HBr (70 kg, 865 mol) was executed.
The mixture was heated at 85 ( 5 °C, for an additional 18
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
209

Benzylic Chloride 14. Alcohol 9a (97.6 kg, 398 mol)
was charged to a 2000 L vessel with heptane (585 L) and
catalytic DMF (1.5 kg), and the resulting slurry was cooled
to 5 ( 5 °C. Thionyl chloride (75.0 kg, 630 mol) was slowly
added over about 1 h. The mixture was stirred at 0-10 °C
until the reaction was complete (about 2 h) upon which time
the reaction was quenched by careful addition of water (500
L). The phases were separated, the aqueous layer was
extracted with heptane (365 L), and the combined organic
layers were washed with a dilute NaOH solution (2 × 400
L, 1 M). The organic solution was filtered through carbon
and reduced in volume under vacuum to provide a solution
of chloride 14, which was carried on as a solution to the
next step without isolation or purification. A sample was
retained, and the solvent was removed under vacuum for
under vacuum at 40-45 °C, for 24 h. Yield: 280.6 kg,
96.8%; mp: 75-76 °C; HPLC: 98.7% (area %); H NMR
1
(400 MHz, DMSO-d6): δ 10.01 (bs, 1H), 7.82-7.79 (m,
2H), 7.71-7.66 (m, 2H), 7.63-7.55 (m, 3H), 7.46-7.42 (m,
2H), 7.03 (d J ) 8.0 Hz, 1H); 13C NMR (100 MHz, DMSO-
d6); δ 141.1; 134.1, 134.0, 133.2, 132.9, 129.3, 128.5, 127.4,
127.0 (q, J ) 4.9 Hz), 126.6, 126.2 (q, J ) 29 Hz), 123.2
(q, J ) 274 Hz); MS m/z 302 (M + H)+, 358 (M + 57)+.
Sulfonamide 6. To a solution of sulfonamide 20 (106.0
kg, 351.8 mol) in THF (562 L) cooled to -5 to -10 °C was
charged a solution of n-butyllithium in hexanes (15% by wt,
330 kg, 773 mol) over several hours, keeping the temperature
below 0 °C throughout the course of addition. After a stir
time of 2 h at <-5 °C the reaction mixture containing
dianion 21 was transferred to a second vessel containing THF
(165 L) into which CO2 was bubbled below the surface of
the THF via a sparging tube. The dianion slurry was
transferred over 3.75 h at <-5 °C with constant addition of
CO2(g) (53.9 kg, 1.23 kmol) during the transfer. After stirring
at <-5 for 2 h, water (225 L) was carefully added, and the
temperature was adjusted to 20-30 °C. The organic solvent
was removed by distillation at atmospheric pressure and the
reaction mixture diluted with water (865 L). The mixture
was extracted with MTBE (2 × 114 kg), and the MTBE
extracts were back extracted with water (145 L), discarding
the MTBE phase. The combined aqueous layers were
acidified with HCl (75 kg, 37% HCl, 760 mol) and extracted
with MTBE (230 kg). The MTBE was reduced in volume
by approximately 75% by distillation at atmospheric pressure.
Acetic anhydride (144 kg, 1410 mol) and methanesulfonic
acid (1.1 kg) were added, and the distillation was continued
as the mixture was heated to 115-125 °C. After holding at
this temperature for 3 h the mixture was cooled to 45 °C,
and methanol (170 L) was added slowly (very exothermic!)
to the reaction mixture. The slurry was cooled to 5 °C and
the product collected by centrifugation, washed sequentially
with methanol (40 L) and MTBE (20 kg), and dried under
vacuum at 50-55 °C for 24 h). Yield: 103.2 kg (89.6%,
for two steps from 20); mp: 203-204 °C; HPLC: 98.9%
(area %); 1H NMR (400 MHz, DMSO-d6): δ 8.45-8.43 (m,
1H), 8.23-8.21 (m, 1H), 8.16-8.12 (m, 1H), 8.08-8.03 (m,
2H), 77.98-7.92 (m, 1H), 7.90-7.86 (m, 1H), 7.78 (d, J )
7.8 Hz, 1H); 13C NMR (100 MHz, DMSO-d6) δ 158.5, 137.1,
136.4, 135.6, 134.3, 133.5, 132.0, 129.6 (q, J ) 31 Hz), 128.5
(q, J ) 4.2 Hz), 125.8, 125.7, 125.3, 122.8 (q, J ) 274 Hz),
122.2; MS m/z 328 (M + H)+, 384 (M + 57)+.
1
analysis. HPLC: 99.8% (area %); H NMR (400 MHz,
DMSO-d6) δ 7.07 (d, J ) 1.8 Hz, 1H), 6.88 (d, J ) 1.8 Hz,
1H), 6.01 (s, 2H), 5.13 (d, J ) 6.8 Hz, 1H), 1.82 (d, J ) 6.8
Hz, 3H); 13C NMR (100 MHz, DMSO-d6) δ 148.5, 143.8
125.3, 122.3, 113.2, 111.9, 102.0, 52.0, 24.8; MS m/z 227/
229 (M - 35)+.
Bromide 10. To a 2000 L reactor were charged NaBH4
(22.0 kg, 582 mol), ZnCl2 (38.7 kg, 284 mol), and heptane
(220 L); the slurry was stirred at 10 ( 10 °C while THF
(450 L) was carefully added (exothermic!). The slurry was
heated at 55 ( 5 °C for about an hour, upon which time a
solution of chloride 14 (in THF, 75 L, and residual heptane)
was transferred over about 30 min. The reaction mixture was
heated at 70 ( 5 °C for about 3 h until all starting material
was consumed. The reaction mixture was cooled to 40 °C
and quenched by addition of water (400 L). The organic
solvent was reduced by vacuum distillation and the residue
extracted with heptane (2 × 175 kg). The product solution
was filtered through carbon and distilled under vacuum,
collecting the fractions that distilled at 65-149 °C at 5-8
mmHg. The isolated oil (71.3 kg) was stirred in a mixture
of hexane (670 L), MeCN (84 L), and water (8 L) for 30
min. The lower MeCN/water layer was discarded and the
organic solvent removed under vacuum to provide bromide
10 as a light-yellow oil. Yield: 63.7 kg, (69.9% from alcohol
1
9a); HPLC purity: 98.4% (area %); H NMR (400 MHz,
DMSO-d6) δ 6.95 (d, J ) 2.0 Hz, 1H), 6.88 (d, J ) 2.0 Hz,
1H), 6.02 (s, 2H), 2.49 (q, J ) 7.4 Hz, 1H), 1.12 (t, J ) 7.4
Hz, 3H); 13C NMR (100 MHz, DMSO-d6) δ 147.9, 144.6,
126.4, 124.1, 112.1, 109.5, 101.5, 21.8, 13.6; MS m/z 228/
230 (M+).
Sulfonamide 20. To a solution of 2-(trifluoromethyl)-
aniline (160.0 kg, 993 mol) and pyridine (97.0 kg, 1226.3
mol) in acetonitrile (78 kg) preheated to 35 ( 5 °C was added
neat benzenesulfonyl chloride (170 kg, 962.5 mol) over about
2 h, maintaining a temperature of 35-45 °C. The mixture
was then heated for 4.5 h at 45-55 °C and cooled to <10
°C. The reaction mixture was quenched by slow addition to
a solution of HCl(aq) (66.0 kg 37% HCl, 1810 mol, in 660
L water) containing 10 g of seed crystals, maintaining a
temperature <10 °C during transfer. The resulting slurry was
stirred overnight (16 h) at <10 °C. The product was collected
by centrifugation, washed with water (300 L), and dried
Diester 26. To a 2000 L vessel were charged sulfonamide
6 (195.0 kg; 596 mol), methanol (488 L), and sodium
methoxide in MeOH (25 wt % in MeOH; 138 kg, 639 mol),
and the mixture was heated to reflux for about 2 h. The
solution was cooled to about 55 °C, whereby methyl
bromoacetate (107.3 kg, 701 mol) was charged slowly,
keeping the temperature of the mixture below 60 °C during
the addition. The reaction mixture was heated at reflux until
analysis indicated that the quantity of monoester 27 in the
mixture was <0.5% by HPLC analysis. The mixture was
cooled to <5 °C and diluted with MTBE (200 L), and the
mixture was recooled to 5 °C. The product was collected by
210
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

centrifugation and washed sequentially with MTBE (125 L)
and water (130 L). The wet cake was reslurried at ambient
temperature in water (500 L) (to remove excess salts not
removed by washing the cake on the centrifuge) and collected
by centrifugation. The product cake was washed with water
(130 L) and dried under vacuum at 50-55 °C for 24 h.
Yield: 216.2 kg (84.1%); mp: 132-133 °C; HPLC: 99.7%
(area %); 1H NMR (400 MHz, DMSO-d6) δ 7.81-7.78 (m,
2H), 7.77-7.73 (m, 2H), 7.72-7.65 (m, 3H), 7.61-7.59 (m,
1H), 4.79 (d, J ) 18.5 Hz, 1H), 4.24 (d, J ) 18.5 Hz, 1H),
3.67 (s, 3H), 3.54 (s, 3H); 13C NMR (100 MHz, DMSO-d6)
δ 168.6, 167.5, 136.4, 133.2, 133.1, 133.0, 132.5, 130.6,
129.7, 129.2, 128.4 (q, J ) 29.9 Hz), 128.0, 127.7 (q, J )
5.5 Hz), 122.9 (q, J ) 274 Hz), 52.6, 52.4, 51.7; MS m/z
432 (M + H)+.
pressure until approximately 400 L remained. At this point
the solution of ketoester 28 can be used directly in the next
step without isolation, as described in the preparation of
triflate 3. To isolate ketoester 28, heptane (210 kg) was
charged, and distillation was continued until a batch tem-
perature of 70 °C was reached. The mixture was cooled to
5 ( 5 °C and the product collected by centrifugation, washed
with heptane (48 kg), and dried for 16 h in a vacuum oven
at 50 ( 5 °C. Yield: 61.5 kg (98.5%) of a yellow solid;
1
mp: 191-192 °C; HPLC: 98.5% (area %); H NMR (400
MHz, DMSO-d6) δ 11.80 (bs, 1H), 8.23-8.21 (m, 1H),
8.00-7.96 (m, 1H), 7.91-7.87 (m, 1H), 7.83-7.81 (m, 2H),
7.57-7.46 (m, 2H), 6.78 (d, J ) 7.8 Hz, 1H), 3.65 (s, 3H);
13C NMR (100 MHz, DMSO-d6) δ 167.5, 157.3, 135.4,
135.3, 133.6, 133.5, 133.4, 129.5, 128.5 (q, J ) 4.2 Hz),
128.3 (q, J ) 30.2 Hz), 127.4, 127.0, 126.6, 123.6 (q, J )
273.8 Hz), 123.3, 109.5, 52.7.
Alternate Preparation of Ketoester 27. To a 250 mL
three-neck round-bottom flask were charged diester 26 (20.25
g, 46.9 mmol), MgCl2 (9.8 g, 103 mmol), and acetonitrile
(75 mL). With vigorous agitation DBU (21.4 g, 140.6 mmol)
was added, keeping the temperature at 40 ( 5 °C. The
reaction mixture was heated to 50-55 °C, and upon
completion of the reaction the acetonitrile was reduced via
vacuum distillation to obtain a thick slurry. The reaction was
quenched with 1 M HCl (140 mL) and the resulting slurry
stirred at 0 ( 5 °C for 1 h. The product was filtered, washed
with cold water, and dried in a vacuum oven at 50 °C to
yield 16.8 g (89.7%) of a yellow solid. HPLC (area %):
98.5%.
Alternate Preparation of Diester 26. To a mixture of 6
(66.0 g, 0.20 mol) and methanol (100 mL) was charged
NaOMe, 25 wt % in methanol (50 mL, 0.22 mol, 0.93 equiv).
The resulting mixture was heated to reflux and held for 4-6
h. Upon completion of the reaction, the mixture was cooled
to room temperature and quenched with 1 M HCl (330 mL).
The product cake was isolated in two crops upon filtering
and washing with water. The two cakes were combined and
dried in a vacuum oven at 50 °C to yield 66.0 g (90.5%) of
monoester 27 as a white solid; mp: 123.7-124.9 °C;
1
HPLC: 98% (area %); H NMR (400 MHz, DMSO-d6) δ
9.64 (s, 1H), 7.93-7.90 (m, 1H), 7.79-7.71 (m, 5H), 7.65-
7.61 (m, 1H), 7.49-7.45 (m, 1H), 7.20 (d, J ) 7.8 Hz, 1H),
3.71 (s, 3H); 13C NMR (100 MHz, DMSO-d6) δ 167.4, 138.4,
133.7, 133.3, 133.0, 131.5, 131.4, 129.2, 128.9, 128.4, 127.4,
126.8 (q, J ) 4.9 Hz), 125.3 (q, J ) 29.5 Hz), 123.2 (q,
J ) 273 Hz), 52.9; MS m/z 360 (M + H)+.
To a clean, dry, 1 L three-neck round-bottom flask were
charged monoester 27 (66.3 g, 0.185 mol), Na2CO3 (45.0 g,
0.423 mol, 2.3 equiv), NaI (6.0 g, 0.04 mol, 0.2 equiv),
acetonitrile (340 mL), and methyl bromoacetate (42.2 g,
0.276 mol, 1.5 equiv). The resulting mixture was stirred and
heated to reflux under nitrogen until completion of the
reaction. The reaction was quenched with water (350 mL)
and the acetonitrile removed under vacuum to provide a
slurry of 26. To the resulting slurry was charged heptane
(100 mL), and the slurry was chilled to 0 ( 5 °C and held
for ∼1 h. Upon filtration, the cake was washed with water
and then heptane and dried in a vacuum oven at 50 °C.
Yield: 78.9 g of diester 26, 99.5% yield (90% for the two
steps) >98% purity by HPLC (area %).
Ketoester 28. To a 2000 L reactor were charged diester
26 (67.5 kg; 156.5 mol) and CH2Cl2 (736 kg), and the
solution was cooled to -30 ( 10 °C. TiCl4 (67.5 kg,
356 mol) was charged and the reaction mixture cooled to
-60 ( 10 °C. Triethylamine (47.5 kg, 469 mol) was added
over about 30 min, maintaining the temperature range. After
stirring for 2-3 h at -60 ( 10 °C, the reaction mixture
was warmed to -15 ( 10 °C and quenched by careful
addition of 3 M HCl(aq) (415 L). The layers were separated,
and the aqueous phase was back extracted with CH2Cl2 (180
kg). The combined organic extracts were washed with water
(210 L) and brine (175 L) and distilled at atmospheric
Triflate 3. A solution of ketoester 28 (in 660 L CH2Cl2,
156.5 mol maximum) was dried over MgSO4 (25.0 kg) until
water by K-F analysis was under 0.1%. The solution was
filtered into a 2000 L vessel and cooled to 0 °C. Pyridine
(31 kg, 493 mol) was charged and the mixture cooled to
-15 ( 5 °C. Trifluoromethanesulfonic anhydride (54.3 kg,
192.5 mol) was added, neat, keeping temperature <0 °C
during the addition. The reaction was quenched after 2 h by
transfer to a separate reactor containing HCl (aq) (23 kg 37%
HCl in 215 L water). The contents were vacuum distilled to
remove most CH2Cl2; the slurry was diluted with heptane
(55 kg) and stirred and cooled to 5 °C. The crude 3 was
collected by centrifugation, washed with water and heptane,
and then recrystallized, solvent-wet, from IPA and MeCN.
The product was isolated by centrifugation and dried in a
vacuum oven at 50 ( 5 °C for 24 h. Yield: 77.5 kg (93.2%
for two steps); mp: 182.5-185.0 °C HPLC: 99.8% (area
1
%); H NMR (400 MHz, DMSO-d6) δ 8.13-8.05 (m, 2H),
7.99-7.96 (m, 2H), 7.89-7.87 (m, 1H), 7.71-7.69 (m, 2H),
7.32-7.30 (m, 1H), 3.67 (s, 3H); 13C NMR (100 MHz,
DMSO-d6) 159.4, 138.3, 134.0, 133.7, 133.5, 132.7, 132.0,
131.0, 128.4 (q, J ) 4.9 Hz), 127.5 (q, J ) 30.2 Hz), 127.4,
126.0, 125.4, 122.9 (q, J ) 274.0 Hz), 122.8, 117.8 (q, J )
321.5 Hz), 53.5; MS m/z 532 (M + H)+.
Ester 29f. To a 1 L flask were charged ketoester 28 (147.8
g, 0.37 mol) and acetonitrile (475 mL). Agitation was started,
and pyridine (87.8 g, 1.1 mol) was added via an addition
funnel as the reaction mixture was warmed to 55 ( 5 °C. A
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
211

solution of 4-fluorobenzenesulfonyl chloride (79.7 g, 0.41
mol) in acetonitrile (150 mL) was added over 20 min. The
reaction mixture was heated at 50-60 °C for 6-18 h, then
cooled to ambient temperature, and slowly transferred to a
2 L flask containing 0.6 M HCl (1 L). After cooling to 0-10
°C and stirring for 1-2 h the precipitate was filtered, washed
with water (2 × 50 mL), and dried in a vacuum oven. The
crude sulfonate ester 29f was charged to a 2 L flask with
isopropyl alcohol (1.2 L) and acetonitrile (300 mL) and
heated to reflux to dissolve. The mixture was filtered then
heated at reflux as water (100 mL) was slowly added. The
reaction mixture was reduced in volume under atmospheric
pressure, collecting 400-450 mL of distillate, then slowly
cooled to 0-5 °C. The product was collected by filtration,
washed with isopropyl alcohol (2 × 40 mL), and dried in a
vacuum oven at 35 ( 10 °C for 18 h. Yield: 173.8 g
(84.3%); mp: 165-166 °C; HPLC purity: 97.1% (area %);
1H NMR (400 MHz, DMSO-d6) 7.98-7.94 (m, 2H), 7.85-
7.82 (m, 2H), 7.80-7.78 (m, 3H), 7.69-7.67 (m, 2H), 7.50-
7.45 (m, 2H), 7.26 (d, J ) 7.0 Hz, 1H). 13C NMR (100 MHz,
DMSO-d6) δ 167.3, 164.7, 159.4, 137.2, 133.5, 133.0, 132.6,
132.3, 132.1, 132.0, 131.4, 130.8, 129.1, 128.7 (d, J ) 95
Hz), 128.5 (d, J ) 95 Hz), 127.8, 127.5 (q, J ) 30.2 Hz),
127.4, 125.8, 122.8 (q, J ) 274 Hz), 122.1, 117.3, 117.1,
52.9; MS m/z 558 (M + H)+.
Boronic Acid 2a. A 2000 L reactor was charged with
THF (284 L) and cooled to -65 ( 10 °C. A solution of
n-butyllithium in hexanes (15% by wt, 86.2 kg, 201.8 mol)
was charged slowly, keeping the temperature below -55 °C.
Bromide 10 (42.0 kg, 183.3 mol) was added neat, keeping
the temperature at -65 ( 10 °C during the addition. After
30 min at this temperature triisopropyl borate (41.2 kg, 219.1
mol) was added neat at -65 ( 10 °C. The mixture was
warmed to -10 ( 10 °C and quenched by addition of a
solution of acetic acid (13.1 kg, 218.2 mol) in water (210
L). The mixture was stirred under nitrogen for 30 min and
then held for the reaction with triflate 3 in the preparation
of ester 30. A small sample was removed for characterization.
HPLC purity: 95.6% (area %); 1H NMR (400 MHz, CDCl3)
δ 7.58 (s, 1H), 7.44 (s, 1H), 6.00 (s, 2H), 2.69 (q, J ) 7.7
Hz, 2H), 1.30 (t, J ) 7.7 Hz, 3H); 13C NMR (100 MHz,
CDCl3) δ 148.9, 146.7, 130.7, 124.9, 111.9, 100.5, 22.4, 13.7;
MS m/z: 195 (M + H)+, 251 (M + 57)+.
79.5 kg of yellow solid (94.6%); mp: 173-174 °C; HPLC
purity: 99.4% (area %); H NMR (400 MHz, CDCl3) δ
1
7.91-7.89 (m, 1H), 7.73-7.71 (m, 1H), 7.61-7.48 (m, 5H),
7.32-7.26 (m, 1H), 6.71 (s, 1H), 6.67 (s, 1H), 6.01 (s, 2H),
2.61 (q, J ) 7.7 Hz), 1.12 (t, J ) 7.7 Hz); 13C NMR (100
MHz, CDCl3) δ 162.5, 147.1, 145.5, 133.7, 133.6, 132.5,
132.2, 131.3, 129.8, 129.7, 129.4 (q, J ) 32.0 Hz), 128.9,
127.7, 127.6, 127.6, 125.3, 123.9, 122.8 (q, J ) 274 Hz),
122.0, 108.3, 101.1, 52.1, 22.5, 14.0; MS m/z 532 (M +
H)+, 588 (M + 57)+.
Alternate Preparation of Ester 30. A 500 mL flask was
charged with 29f (56.8 g, 102 mmol), Na2CO3 (13.5 g, 131
mmol), PdCl2(PPh3)2 (106 mg, 0.24 mmol), and triphen-
ylphosphine (132 mg, 0.50 mmol). The flask was swept with
N2, and a boronic acid solution (approximately 116 mmol,
prepared as previously described) was transferred to the flask.
The flask was covered with foil to protect from light and
pressure-purged with nitrogen (10×) with stirring. The
reaction mixture was heated at reflux for about 48 h until
all starting material was consumed. The mixture was cooled
to 55-60 °C, 50% NaOH (4.0 g) and trithiocyanuric acid
(200 mg) were added, and the mixture was heated at 60-65
°C for about 30 min. The mixture was cooled to 55 °C and
transferred to a separatory funnel, discarding the lower
aqueous phase. The organic phase was returned to the flask
and reheated to above 55 °C, whereby isopropyl alcohol (75
mL) and then water (60 mL) were added in sequence,
keeping the temperature above 55 °C. The reaction mixture
was heated at 55 °C for 1 h and then slowly cooled to 0-5
°C. The mixture was held at this temperature for 1 h, filtered,
washed with IPA (15 mL), and dried in a vacuum oven at
40 ( 10 °C overnight. Yield: 47.1 g, 86.9%; HPLC purity:
99.8%.
CI-1034 (1). To a 2000 L reactor were charged ester 30
(79.3 kg, 149.2 mol), KOH (10.1 kg, 91%, 163.8 mol),
2-butanol (640 kg), and water (67 L); the mixture was heated
at 75 ( 5 °C for about 4 h. After filtering through carbon
and a polish filter, the volume was reduced to approximately
600 L by atmospheric distillation. The contents were cooled
to -5 ( 5 °C and held in this range for about 1 h. The
product was collected by centrifugation, washed with cold
2-butanol, and dried in a vacuum oven at 40-50 °C for 24
h. Yield: 64.9 kg, 78.3%. HPLC purity: 99.68% (area%);
1
Ester 30. A 1600 L reactor was charged with triflate 3
(84.0 kg, 158.1 mol), Na2CO3 (12.6 kg, 118.9 mol), PdCl2-
(PPh3)2, (277 g, 0.39 mol), PPh3 (208 g, 0.79 mol); this
mixture was pressure-purged with nitrogen (3×). The solu-
tion of boronic acid 2a was then transferred to the reactor,
and the contents were heated to 65 ( 5 °C until the reaction
was complete (overnight). The mixture was treated with 50%
NaOH (6 kg) and trithiocyanuric acid (0.55 kg, 3.1 mol) and
stirred for 30 min. The lower aqueous layer was removed,
the reaction mixture was diluted with IPA (185 L) and water
(250 L), and the temperature was adjusted to 50 ( 5 °C and
held for 1 h. The mixture was cooled to 0 ( 5 °C, and ester
30 was collected by centrifugation, washed with IPA, and
dried in a vacuum oven at 40 ( 10 °C overnight. Yield:
[K+] ) 7.22% (theory ) 7.04%); H NMR (400 MHz,
DMSO-d6) δ 7.76-7.69 (m, 3H), 7.64-7.60 (m, 2H), 7.57-
7.53 (m, 1H), 7.46-7.42 (m, 1H), 7.02 (d, J ) 8.0 Hz), 6.79
(s, 1H), 6.63 (s, 1H), 2.50 (m, 2H), 1.16 (t, J ) 7.7 Hz); 13
C
NMR (100 MHz, DMSO-d6) δ 162.5, 146.1, 143.9, 143.7,
136.0, 133.7, 133.6, 133.6, 132.5, 131.7, 130.9, 130.3, 129.5,
128.8 (q, J ) 30.2 Hz), 126.7 (q, J ) 4.9 Hz), 126.4, 126.1,
124.4, 123.7, 123.0 (q, J ) 274.5 Hz), 120.8, 111.6, 109.5,
100.6, 22.2, 13.8; MS (of the free acid) m/z 518 (M + H)+,
574 (M + 57)+.
Received for review July 27, 2003.
OP034104G
212
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development