Full text of DOI:10.1021/op100268e

Organic Process Research & Development 2011, 15, 148–157
Development of an Early Enabling Synthesis for PF-03052334-02: A Novel
Hepatoselective HMG-CoA Reductase Inhibitor
Daniel M. Bowles,* David C. Boyles, Chulho Choi, Jeffrey A. Pfefferkorn, and Stephanie Schuyler
Pfizer Global Research and DeVelopment, Groton Laboratories, MS8156-064 Eastern Point Road, Groton,
Connecticut 06340, United States
Edward J. Hessler
Bridge Organics, 311 West Washington Street, Vicksburg, Michigan 49097, United States
Abstract:
Early process development work toward a promising pyrazole-
based HMG-CoA reductase inhibitor is described. PF-03052334-
02 (1) was prepared in 14 synthetic steps with a 21% overall yield,
highlighted by a modified three-step hydroxypyrazole formation
in which the yield was improved from 37% to 73%, a Suzuki/
ozonolysis pathway that streamlined the downstream chemistry,
and a reversed Wittig olefination strategy that improved the key
Figure 1. HMG-CoA reductase inhibitor candidate PF-
coupling step from 50% to 95% yield. Multiple process hazards
and most chromatography steps were removed, and a highly
effective active pharmaceutical ingredient (API) salt formation,
purification, and isolation protocol was also developed.
03052334 (1).
First-Generation Route. While the original medicinal
chemistry route (Scheme 1)5 was effective at generating a
variety of analogues, it was relatively linear, and as a result the
throughput was somewhat low (5.6% yield over 15 linear
steps).7 The synthesis also contained several nonprocess-friendly
steps by using diazonium chemistry, a nitrilimine formation, a
1,3-dipolar cycloaddition reaction that afforded a regioisomeric
mixture, ceric ammonium nitrate oxidation, and a tedious
borane/Dess-Martin reduction/oxidation sequence all in order
to reach intermediate pyrazole aldehyde 11.
In addition, the downstream conversion of aldehyde 11 to
1, while capable of supporting the manufacture of subgram
active pharmaceutical ingredient (API) quantities, was not
convenient for further scale-up (Scheme 2). NaBH4 reduction
of amide 14 was followed by treatment with Ph3P•HBr, and
resulting ylide 16 was treated with costly and unstable aldehyde
178 to afford only a 50% yield of 18 as a mixture of cis/trans
isomers.9 The mixture was deprotected, reduced, and saponified
to produce 1 as an amorphous foam that was contaminated with
inorganic impurities.
Introduction
Statins, 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-
CoA) reductase inhibitors,1 are regularly prescribed in the battle
against coronary heart disease (CHD).2,3 While statins are
generally well tolerated, musculoskeletal side effects such as
myalgia can limit compliance with treatment regimens; more-
over, in rare cases these musculoskeletal effects can progress
to a serious and life-threatening condition known as rhabdomy-
olysis.4 During ongoing studies, our Discovery Chemistry
colleagues became interested in pyrazole PF-03052334-02 (1,
Figure 1), an HMG-CoA reductase inhibitor5 characterized by
promising efficacy and high hepatoselectivity that may lead to
reduced risk of musculoskeletal effects.6
* To
whom
correspondence
should
be
addressed.
E-mail:
dan.bowles@pfizer.com.
(1) McKenney, J. M. Clin. Cardiol. (Suppl. III). 2003, 26, 32–38.
(2) Grundy, S. M. J. Intern. Med. 1997, 241, 295–306.
Further complicating the early route, nearly every step
required cleanup by silica gel chromatography. As a result, the
original route totaled >25 operational steps when considering
scale up into our kilo-lab facility. With an increasing demand
for API in support of early toxicological and clinical studies, it
became critical to develop a second generation route capable
of delivering hundred-gram quantities of 1. In designing such
(3) Jones, P.; Kafonek, S.; Laurora, I.; Hunninghake, D. Am. J. Cardiol.
1998, 81, 582–587.
(4) Cerivastatin was withdrawn from the world market because of its
potential for severe myotoxic effects. For reviews, see: (a) Rosenson,
R. Am. J. Med. 2004, 116, 408–416. (b) Tiwari, A.; Bansal, V.; Chugh,
A.; Mookhtiar, K. Expert Opin. Drug Saf. 2006, 5, 651–666. For
related information see: (c) Graham, D. J.; Staffa, J. A.; Shatin, D.;
Andrade, S. E.; Schech, S. D.; La Grenade, L.; Gurwitz, J. H.; Chan,
K. A.; Goodman, M. J.; Platt, R. JAMA 2004, 292, 2585–2590.
(5) Pfefferkorn, J. A.; Choi, C.; Larsen, S. D.; Auerbach, B.; Hutchings,
R.; Park, W.; Askew, V.; Dillon, L.; Hanselman, J. C.; Lin, Z.; Lu,
G. H.; Robertson, A.; Sekerke, C.; Harris, M. S.; Pavlovsky, A.;
Bainbridge, G.; Caspers, N.; Kowala, M.; Tait, B. D. J. Med. Chem.
2008, 51, 31–45.
(7) Total step count does not include additional operations necessary for
silica gel chromatography cleanup and for the synthesis of Konoike
ylide 32.
(8) Ra´dl, S. Synth. Commun. 2003, 33, 2275–2283.
(9) The mixture of cis/trans isomers was not considered an issue because
the olefin is reduced downstream.
(6) For a review of other recent hepatoselective statins, see: Pfefferkorn,
J. A. Curr. Opin. InVest. Drugs 2009, 10, 245–252.
148
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development
10.1021/op100268e  2011 American Chemical Society
Published on Web 12/17/2010

Scheme 1. First-generation route to pyrazole aldehyde 11
Scheme 2. First-generation endgame
a route, we sought to remove the most hazardous and nonscal-
ablestepswhileeliminatingtheneedforsilicagelchromatography.
were able to improve the three step yield of 24b to 73% by
using diethyl oxalate instead of dibenzyl oxalate.11 Cryogenic
treatment of a THF solution of ethyl isovalerate 20 with LDA
followed by slow addition of 19b afforded oxalate-ester 21b,
which after aqueous workup was treated with 4-fluorophenyl-
hydrazine hydrochloride 22 in aqueous acetonitrile to produce
hydrazone 23b (not isolated). Acetonitrile was exchanged with
refluxing acetic acid, during which time ring closure resulted
in smooth conversion to 24b. The crude product was isolated
by crystallization from toluene/heptane in 73% yield (99%
regioselectivity,12 >99% HPLC purity) in over three steps on
Results and Discussion
Hydroxypyrazole Synthesis. Based on a route previously
developed by our medicinal chemistry colleagues toward a
related oxypyrazole series,10 we were able to readily synthesize
hydroxypyrazole 24b as a key intermediate for the second
generation synthesis of 1 (Scheme 3).
While the original medicinal chemistry synthesis afforded
only a 32-37% yield of corresponding benzyl ester 24a, we
(10) Larsen, S. D.; Poel, T. J.; Filipski, K. J.; Kohrt, J. T.; Pfefferkorn,
J. A.; Sorenson, R. J.; Tait, B. D.; Askew, V.; Dillon, L.; Hanselman,
J. C.; Lu, G. H.; Robertson, A.; Sekerke, C.; Kowala, M. C.; Auerbach,
B. J. Bioorg. Med. Chem. Lett. 2007, 17, 5567–5572.
(11) Yield improvement is attributed to enhanced stability of the oxalate
to reaction conditions and ease of purification by crystallization.
(12) The >99% regioselectivity was determined by chiral HPLC vs chiral
1
standards and verified by H NMR.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
149

Scheme 3. Synthesis of hydroxypyrazole 24
hydroxyl group in 24b to its corresponding triflate by treatment
with trifluoromethanesulfonic anhydride and triethylamine in
dichloromethane. Following aqueous workup, the crude product
mixture was crystallized from aqueous 2-propanol to provide
26 in 95% yield with >99% HPLC purity on 4.5 kg scale.
Activation and Coupling. As illustrated in Scheme 6, we
found it difficult to couple triflate 26 with complex dihydroxy-
ester boronic acid moieties such as 27; however, we observed
that styrenylboronic acid 29 was an effective coupling partner.
Thus, we investigated a Suzuki/ozonolysis route to pyrazole
aldehyde 11b. Unfortunately, we initially observed that the
Suzuki coupling was not robust upon scale-up.17 However, after
reaction optimization screening we observed that dichlorobis-
(triphenylphosphine)palladium(II) with a 4:1:1 ratio of toluene/
isopropyl alcohol/water was quite effective for this transfor-
mation. A series of bases and additives were also screened, and
the combined use of potassium carbonate with either potassium
bromide or benzyltriethylammmonium chloride significantly
improved conversion to pyrazole 30, which was conveniently
crystallized and filtered from the crude reaction mixture in 91%
yield (>98% purity) on 3.5 kg scale.
21 kg scale. In one pilot batch, cooling was lost during the
oxalate addition, and the reaction temperature reached -30 °C.
As a result, the three step process afforded 24b of poor quality
(<50% purity).
Pyrazole Activation. With a steady supply of 24b in hand,
we focused our effort toward conversion of the hydroxyl
functionality into a more suitable handle for dihydroxyester side-
chain attachment.13 A retrosynthetic analysis (Scheme 4) based
on a wealth of literature examples14 highlighted several potential
alternatives to the medicinal chemistry approach (Wittig Route
1) including direct pyrazole alkylation, metal-mediated coupling,
or reversing the Wittig partners (Wittig Route 2). Unfortunately,
the most convergent of these options met with the most
difficulty. In particular, a general lack of activity toward pyrazole
lithiation ruled out a direct alkylation approach,15 and attempts
toward halogenation (formation of 25) followed by metal-
halogen exchange were also not successful (Scheme 5). In
addition, we observed that coupling unsaturated dihydroxyester
side-chain moieties to an activated pyrazole core (Suzuki,
Sonagashira, Heck coupling) were not feasible in our hands
toward substrates such as 28.16 Instead, we converted the
Ozonolysis and Amide Formation. The styrene functional-
ity in 30 (Scheme 6) was readily cleaved in batch mode on
750 g scale by bubbling ozone gas through a cryogenic solution
in dichloromethane and methanol followed by quenching with
dimethylsulfide. Crude workup resulted in isolation of aldehyde
11b in 76% yield.
The resulting crude ester was saponified with sodium
hydroxide in aqueous methanol and was extracted into DCM
to afford 31 in good yield (75%). Activation with EDAC ·HCl/
HOBt followed by treatment with 4-methylbenzylamine in
DMF/DCM at 0 °C formed amide 14, which was isolated by
aqueous workup followed by crystallization from acetonitrile
in 89% yield with >97% HPLC purity on 400 g scale.
Considering the hazardous potential with batch ozonolysis
on scale, we also sought to investigate a flow-ozonolysis
approach to conversion of 30 to 14. Over 1 kg of 30 was treated
with ozone in a flow-manner using a relatively simple home-
made reactor,18 in which a steady stream of ozone was bubbled
through a flowing countercurrent of olefin solution in a mixing
tube. The resulting ozonide solution was directed into a stirring
quench tank with dimethylsulfide in methanol to afford what
was thought to be aldehyde 11b. Unfortunately, the batch of
30 used in this case was isolated differently than previous
batches and was likely contaminated with residual HCl,19 and
as a result dimethylacetal 11c20 was the major ozonolysis
product (Scheme 7) in our flow campaign. Fortunately, the flow-
ozonolysis batch was recovered by saponification followed by
hydrolysis of the acetal with aqueous TFA in DCM to afford
31 in 60% yield on 500 g scale.
`
(13) For a recent review article related to the synthesis of statins, see: Easar,
Z. Curr. Org. Chem. 2010, 14, 816–845.
(14) Additional references from a retrosynthetic approach include: (a)
Watanabe, M.; Koike, H.; Ishiba, T.; Okada, T.; Seo, S.; Hirai, K.
`
Bioorg. Med. Chem. 1997, 5, 437–444. (b) Easar, Z.; Tramsˇek, M.;
Gorsˇek, A. Acta Chim. SloV. 2010, 57, 66–76. (c) Kim, H. S.; Kim,
H.; Sim, J. Y.; Cho, S. M.; Kim, W. J.; Suh, K. H.; Lee, G. S. PCT
Int. Appl. 41, 2010. (d) Wu, X.; Wang, L.; Wang, S.; Chen, Y. Amino
Acids 2010, 39, 305–308. (e) Hobson, L. A.; Akiti, O.; Deshmukh,
S. S.; Harper, S.; Katipally, K.; Chiajen, J.; Livingston, R. C.; Lo, E.;
Miller, M. M.; Ramakrishnan, S.; Shen, L.; Spink, J.; Tummala, S.;
Wei, C.; Yamamoto, K.; Young, J.; Parsons, R. L. Org. Process Res.
DeV. 2010, 14, 441–458. (f) Liljeblad, A.; Kallinen, A.; Kanerva, L. T.
Curr. Org. Synth. 2009, 6, 362–379. (g) Job, A.; Stolle, A. Asymmetric
Synth. 2007, 1, 326–330. (h) Bergeron, S.; Chaplin, D. A.; Edwards,
J. H.; Ellis, B. S. W.; Hill, C. L.; Holt-Tiffin, K.; Knight, J. R.;
Mahoney, T.; Osborne, A. P.; Ruecroft, G. Org. Process Res. DeV.
2006, 10, 661–665.
(15) A wide variety of pyrazole esters and amides were found to be either
unreactive or unstable when treated with alkyllithium bases in our
hands. It is unknown whether alkyllithium bases deprotonated the
methine proton, possibly leading to further decomposition pathways.
(16) We were unsuccessful in converting the hydroxyl functionality to its
halogen counterpart. We continue to study alternate methods for
hydroxypyrazole activation and intend to publish related work in a
future manuscript. In particular, failed attempts toward forming
substrate 28 via Suzuki with activated pyrazoles seemed to result more
from relatively poor stability of a variety of coupling partners similar
to 27 in our hands.
(17) In early Suzuki trials we observed significant decomposition (30-
50%) of the triflate. Optimization of the catalyst, solvent system, base,
and additives reduced decomposition to the corresponding hydroxy-
pyrazole almost entirely.
(18) A similar setup was described by Pelletier and coworkers: Pelletier,
M. J.; Fabiilli, M. L.; Moon, B. Appl. Spectrosc. 2007, 61, 1107–
1115. Flow ozonolysis can readily be outsourced to Bridge Organics,
http://www.bridgeorganics.com.
(19) The batch of olefin used in the flow ozonolysis campaign, unlike the
material for one-pot ozonolysis, was isolated after a dilute HCl wash.
1
(20) As determined by H NMR and MS analysis.
150
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Scheme 4. Retrosynthetic analysis of dihydroxyester side-chain attachment
Scheme 5. Activation of the hydroxypyrazole core via
formation of triflate 26
genated with 10% Pd/C in a Parr reactor to afford 36 in 91%
yield over 2 steps.25
API Isolation. In the original second generation process,
the palladium catalyst was filtered off and the crude product
mixture was treated with 50% NaOH to afford crude carboxy-
late salt 1, which was acidified, purified using chromatography,
treated with caustic, and concentrated to dryness to afford small
quantities of API contaminated with residual NaOH. A more
practical salt formation based on a method previously developed
in our laboratories26 was applied upon scale-up. The crude
sodium salt was first adjusted to pH 5 with acetic acid and then
extracted into ethyl acetate as free acid 37. Removal of ethyl
acetate for2-propanol by distillation at reflux afforded a mixture
of 37 and lactone 38. As a result, a 1:1 sodium salt was easily
formed in a stoichiometric manner by slowly charging aqueous
NaOH until the level of 38 fell to below 0.5% by HPLC
analysis.27 Residual water was removed via azeotropic distil-
lation until a 97:3 ratio of 2-propanol to water was obtained,28
and crude sodium salt 1 was isolated by filtration. Recrystal-
lization, filtration, and drying afforded >160 g of the desired
API in an 81% yield with excellent chemical (>99%) and
stereochemical purity (>99.5% de) with the desired physical
properties.29,30
Side-Chain Coupling and Reduction. In a significant
improvement over the original Wittig strategy (Scheme 2),
aldehyde 14 was readily coupled with stabilized Konoiki ylide
3221 in refluxing toluene to afford 33 a 95% yield of a mixture
of cis/trans olefins with excellent retention of stereochemistry
(Scheme 8, >99.5% ee).22 Unfortunately we observed that the
triphenylphosphine oxide byproduct had a deleterious effect on
downstream chemistry if not sufficiently purged. As we were
unable to isolate the desired product by crystallization, we were
forced to purify the crude reaction mixture on a silica gel pad
to afford 33 in 95% yield on 660 g scale. We were also met
with some difficulty in isolating downstream intermediates, as
most steps leading up to API crystallization resulted in crude
oils.23
Deprotection with aqueous HF in acetonitrile afforded
ꢀ-hydroxy ketone 34 in 93% yield. Reduction with sodium
borohydride24 afforded very high stereoselectivity for the desired
olefin syn-diol diastereomer 35 (>99% de), which was hydro-
(21) Konoike, T.; Araki, Y. J. Org. Chem. 1994, 59, 7849–7854.
(25) With excellent retention of stereochemistry (>99% de).
(26) Bowles, D. M.; Bolton, G. L.; Boyles, D. C.; Curran, T. T.; Hutchings,
R. H.; Larsen, S. D.; Miller, J. M.; Park, W. K. C.; Ritsema, K. G.;
Schineman, D. C. Org. Process Res. DeV. 2008, 12, 1183–1187.
(27) Sodium hydroxide selectively consumes free acid 37, followed by
reaction with lactone 38, in that order. This phenomenon allowed
careful titration to the desired addition endpoint.
1
(22) The relative ratio of E/Z isomers was estimated at 60:40 E/Z by H
NMR analysis, although this result is effectively inconsequential due
to downstream olefin reduction.
(23) Oil intermediates were typically isolated by concentration to dryness
or were carried forward as crude solutions. Any attempts to further
scale this chemistry would benefit greatly from identification of more
solid isolation points after the Wittig chemistry. For one potential
alternative to dealing with late-stage non-crystalline intermediates, see:
Klajic, A.; Zupet, R. PCT Int. Appl. 39, 2010.
1
(28) As indicated by H NMR analysis.
(29) Crystalline form A with rodlike 10-20 µm particle size.
(30) Small pilot batches with varying amounts of isomeric impurities were
found to be upgraded from ∼95-97% de to >99.5% de using this
recrystallization protocol.
(24) Standard syn-reduction conditions: diethylmethoxyborane and sodium
borohydride in 3:1 THF/MeOH at <-70 °C.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
151

Scheme 6. Sukuki/Ozonolysis route to amide-aldehyde 14
Scheme 7. Recovery of the flow-ozonolysis batch
Conclusion
2-Isopropyl-3-oxo-succinic Acid Diethyl Ester (21b). Labo-
ratory Procedure. To a solution of diisopropylamine (1.13 L,
816 g, 8.02 mol) in THF (1.0 L) held below -70 °C was
charged n-butyllithium (2.5 M solution in hexanes, 3.20 L,
512 g, 8.00 mol, 1.0 equiv) at a rate to maintain the pot
temperature below -60 °C. After stirring for 30 min, ethyl
isovalerate 20 (1.06 kg, 8.15 mol, 1.0 equiv) was added
dropwise via addition funnel, maintaining the pot temperature
between -75 to -60 °C. The mixture was stirred at this
temperature for 1 h, after which time diethyl oxalate 19b (1.2
kg, 8.21 mol, 1.02 equiv) was charged via addition funnel,
holding the reaction between -70 to -60 °C. The cold bath
was removed, and the pot mixture was warmed to -20 °C over
2 h, after which time HPLC analysis indicated >99.5%
consumption of 20. The reaction mixture was charged with
glacial acetic acid (814 g, 13.6 mol, 2.26 equiv) followed by
water (1.5 L), and the mixture was allowed to warm to 20 °C
over 2 h. After phase separation, the organic layer was washed
with water (1.5 L), saturated sodium bicarbonate solution (1.5
L), and saturated sodium chloride solution (1.5 L), and was
concentrated at <50 Torr to produce an oil that was held under
vacuum for 2 h to afford 1.79 kg (97% yield, >98% HPLC
purity, RRT 13.6 min) of 21b as an orange oil that was carried
A relatively low-yielding first-generation synthetic route31
that was heavily dependent on nonprocess-friendly chemistry
and silica gel chromatography was replaced with a more scalable
synthesis enabling manufacture of several hundred grams of 1
in a matter of weeks. Utilization of hydroxypyrazole 24b as a
core starting material enabled a rapid synthesis of amide
aldehyde 14, which allowed downstream conversion to 166 g
of sodium salt 1 in support of phase I clinical studies. The
synthesis was completed in 14 steps32 with a 21% overall yield,
including removal of all but one silica gel chromatography step.
Although further process development will be needed before
larger-scale plant campaigns may be realized, the enabling work
presented herein has resulted in rapid advancement of this
promising cardiovascular candidate into early clinical trials.
Experimental Section
HPLC analyses of chemical purity were carried out using
an Agilent 1100 HPLC using a YMC Pack Pro C-18 column
(150 mm × 4.6 mm, 3 µm) with two mobile phases: 0.2%
HClO4 in 95:5 water/acetonitrile (mobile phase A) and aceto-
nitrile (mobile phase B); and a gradient method: 20% eluent B
linearly increased to 95% eluent B over 20 min, 1.0 mL/min,
235 nm. HPLC chiral purity analyses were carried out using
an Agilent 1100 HPLC using a Diacel Chiralpak AD column
(250 mm × 4.6 mm, 10 µm) with an isocratic mobile phase of
hexanes/2-propanol 85:15, 1.0 mL/min, 210 nm.
1
forward without further purification. H NMR (400 MHz,
CDCl3) δ 4.36-4.29 (m, 2H), 4.21-4.15 (m, 2H), 3.84 (d, J
) 6.8 Hz, 1H), 2.52-2.43 (sept, J ) 6.8 Hz, 1H), 1.36 (t, J )
7.2 Hz, 3H), 1.25 (t, J ) 7.1 Hz, 3H), and 1.03-0.99 (m, 6H).
MS (APCI+): 231 [M + H].
Kilogram-Laboratory Procedure (Not Isolated). To a solu-
tion of diisopropylamine (13.7 L, 9.9 kg, 97.8 mol) in THF
(16 L) held between -75 to -70 °C was charged n-butyllithium
(31) A 5.6% yield over 15 linear steps (23 steps overall including ylide
synthesis) without including multiple silica gel chromatography
operations.
(32) Excluding the synthesis of side-chain ylide 32, which remained
virtually unchanged from the original first-generation route.
152
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Scheme 8. Side-chain attachment and deprotection
(2.5 M solution in hexanes, 39.1 L, 27.1 kg, 97.8 mol, 1.0
equiv), holding the pot temperature < -60 °C.33 After stirring
for 30 min, ethyl isovalerate 20 (12.9 kg, 99.1 mol, 1.0 equiv)
was charged via transfer pump at < -60 °C during the addition.
The mixture was stirred between -75 to -65 °C for 1 h, after
which time diethyl oxalate 19b (14.5 kg, 99.3 mol, 1.01 equiv)
was charged via transfer pump at < -60 °C. The reactor jacket
temperature was raised to warm the pot mixture to -20 °C
over 2 h, after which time HPLC analysis indicated >99.5%
consumption of 20. The reaction mixture was transferred into
a 10 °C solution of glacial acetic acid (13.2 kg, 12.6 L, 220
mol, 2.26 equiv) and water (80 L), and the mixture was warmed
to 20 °C after addition was complete. The phases were
separated, and the organic layer was washed with water (25
L), saturated sodium bicarbonate solution (20 L), and 5 wt %
sodium chloride solution (20 L). The resulting mixture was
concentrated by distillation at 40 °C under vacuum to 20 L.
Acetonitrile (30 L) was charged, and the resulting mixture was
distilled at 40 °C and 40 Torr to isolate crude 21b as a solution
in acetonitrile, final volume 30 L, that was carried forward into
the second step without further purification (assumed yield
100%). A small aliquot was concentrated and was found to
exhibit the desired representative spectral properties listed above.
2-[(4-Fluorophenyl)-hydrazono]-3-isopropyl-succinic Acid
Diethyl Ester (23b). Laboratory Procedure. To a slurry of
4-fluorophenylhydrazine hydrochloride 22 (565 g, 3.48 mol,
1.12 equiv) in water (800 mL) was charged a solution of 21b
(714 g, 3.10 mol, 1.0 equiv) in acetonitrile (1.0 L). The biphasic
suspension was heated to between 45-50 °C for 24 h and the
reaction mixture was separated. The aqueous layer was diluted
with water (2 L) and extracted with MTBE (2 × 2 L). The
combined organic layers were washed with saturated brine
solution (2 × 2 L), concentrated to dryness at <50 Torr, and
the resulting oil was dried at 20 °C and 10 Torr for 16 h to
afford hydrazone 23b (1.02 kg, 99% yield, 96% HPLC purity,
RRT 7.8 min), as an orange oil. 1H NMR (400 MHz, CDCl3)
δ 7.14-7.10 (m, 2H), 7.00-6.96 (m, 2H), 4.25 (q, J ) 7.2
Hz, 2H), 4.14 (q, J ) 7.2 Hz, 2H), 3.45 (d, J ) 8.2 Hz, 1H),
2.50 (sept, J ) 6.8 Hz, 1H), 1.32 (t, J ) 7.1 Hz, 3H), 1.23 (t,
J ) 7.1 Hz, 3H), 1.05 (d, J ) 6.8 Hz, 3H), and 0.98 (d, J )
6.7 Hz, 3H). 19F NMR (376 MHz, CDCl3) δ -122.40. MS
(APCI+): 339 [M + H]
Kilogram-Laboratory Procedure. To a slurry of 4-fluo-
rophenylhydrazine hydrochloride 22 (18.4 kg, 113 mol,
1.16 equiv) in water (26.0 L) was charged a crude solution
of 21b in acetonitrile (carried forward without isolation
from step 1; assumed 100% yield, 97.8 mol) in acetonitrile,
total volume 30 L. The biphasic suspension was stirred
(33) Approximately 30 min.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
153

at between 45-50 °C for 24 h and the reaction layers
were separated. The aqueous layer was diluted with water
(65 L) and extracted with MTBE (2 × 65 L). The
combined organic layers were washed with 10 wt % brine
solution (2 × 65 L) and concentrated by vacuum distil-
lation at 30 °C and 60 Torr to a final volume of 30 L to
afford hydrazone 23b as a crude solution in MTBE (yield
assumed to be 100%) that was carried forward without
further purification. A small aliquot was found to exhibit
the desired representative spectral properties listed above.
1-(4-Fluorophenyl)-5-hydroxy-4-isopropyl-1H-pyrazole-
3-carboxylic Acid Ethyl Ester (24b). Laboratory Procedure.
A solution of hydrazone 23b (1.27 kg, 3.75 mol, 1.0 equiv) in
glacial acetic acid (2.0 L) was heated to reflux for 16 h, cooled
to 80 °C, and concentrated to dryness at <30 Torr. The resulting
solids were reconcentrated from toluene (3 × 750 mL) on the
rotary evaporator to remove residual acetic acid. The resulting
brown solids were recrystallized from a mixture of toluene (1.8
L) and n-heptane (3.6 L), rinsed with n-heptane (1 × 2 L), and
dried for 16 h at 50 °C to provide hydroxypyrazole 24b (0.88
kg, 80% yield, >99% HPLC purity, RRT 8.9 min) as a white
over 1.5 h at < -20 °C. The reaction mixture was warmed to
18-25 °C and was stirred for 8 h, until less than 0.5% 24b
was observed by HPLC. The reaction mixture was quenched
with water (2 L) and the layers were separated. The lower
organic layer was extracted with HCl (1N, 2 L) followed by
water (2 L). The resulting dichloromethane solution was charged
with 2-propanol (7 L) and concentrated on the rotary evaporator
to remove dichloromethane. The resulting solution was heated
to 60 °C, charged with water (6 L) at >50 °C, then cooled at a
rate of 0.5 °C/min to 5 °C. The resulting solids were filtered,
washed with aqueous 2-propanol (3:1 water/IPA, 2 × 1.6 L),
and dried for 48 h at 45 °C to afford 26 (4.4 kg, 95% yield,
>99% HPLC purity, RRT 11.3 min) as a white granular solid.
1H NMR (400 MHz, CDCl3) δ 7.52-7.45 (m, 2H), 7.21-7.17
(m, 2H), 4.49-4.40 (q, 2H), 3.32-3.21 (m, 1H), 1.43-1.38
(m, 9H). 13C NMR (100 MHz, DMSO-d6) δ 14.90, 21.14,
24.11, 61.48, 101.15, 117.37, 117.45, 120.30, 128.01, 137.29,
140.52, 160.92, 164.91. MS (APCI+): 425 [M + H]. Anal.
Calcd for C16H16F4N2O5S: C, 45.28; H, 3.80; N, 6.60; Obs: C,
45.26; H, 3.91; N, 6.59.
1
1-(4-Fluorophenyl)-4-isopropyl-5-styryl-1H-pyrazole-3-
carboxylic Acid Ethyl Ester (30). To a nitrogen-purged slurry
of triflate 26 (3.5 kg, 8.25 mol), styrenylboronic acid 29 (1.31
kg, 8.8 mol, 1.07 equiv), potassium bromide (196 g, 1.65 mol,
0.20 equiv), and potassium carbonate (3.4 kg, 24.7 mol, 3.0
equiv) was charged a mixture of toluene (24 L), water (7 L),
and 2-propanol (7 L), and the resulting slurry was treated with
dichloro-bis(triphenylphosphine)palladium(II) (145 g, 207 mmol,
2.5 mol %), and heated to reflux for 12 h. The phases were
separated, and the product containing organic layer was
extracted in succession with potassium carbonate (2 N, 2.5 L),
water (2.5 L), and HCl (3 N, 2.5 L).35 The solvent mixture was
switched to 2-propanol via continual distillation at ambient
pressure, affording a crude solution of 30 in 2-propanol (5 L).
The solution was cooled to 50 °C, clarified by filtration through
a 0.2 µm cartridge, and was charged with water (1 L). The
resulting solution was cooled at a rate of 1 °C/min to 5 °C with
stirring over 1 h and the resulting solids were filtered, washed
with 2-propanol/water (2:1, 3 L) and dried at 45 °C and 20
Torr for 12 h to afford 30 (2.8 kg, 91% yield, >99.5% HPLC
to off-white solid. H NMR (400 MHz, CDCl3) δ 7.69-7.65
(m, 2H), 7.34-7.30 (m, 2H), 4.24 (q, J ) 7.1 Hz, 2H), 3.27
(sept, J ) 7.0 Hz, 1H), 1.41 (t, 3H), 1.28-1.22 (m, 6H). 13C
NMR (100 MHz, DMSO-d6) δ 14.31, 22.15, 23.98, 60.54,
116.11, 116.26, 125.54, 135.18, 139.99, 149.83, 160.12, 163.79.
19F NMR (376 MHz, CDCl3) δ -115.28. MS (APCI+): 293
[M + H]. Anal. Calcd for C15H17FN2O3: C, 61.63; H, 5.86; N,
9.58; Obs: C, 61.60; H, 5.89; N, 9.52.
Kilogram-Laboratory Procedure. To a solution of hydrazone
23b in MTBE (total volume 30 L, assumed to be 100% yield
from steps 1-2, 97.8 mol 23b) was charged glacial acetic acid
(52 L), and the mixture was distilled at 30 Torr and 50 °C until
MTBE had been fully displaced. The resulting solution was
heated to reflux for 20 h until 23b had been fully consumed
(less than 0.5% 23b by HPLC analysis). The pot mixture was
distilled at 30 Torr and 60 °C to a total volume of 35 L. Toluene
(20 L) was added, and the pot was redistilled to 35 L. The
toluene charge and distill cycle was repeated twice more to
afford a crude solution of 24b in toluene (35 L total volume).
The pot mixture was charged with toluene (40 L), heated to 50
°C, and charged with n-heptane (94 L). The resulting slurry
was heated to 80 °C to fully dissolve any solids and then cooled
at a rate of 1 °C/min to 10 °C. The resulting solids were
collected by filtration, washed with n-heptane (2 × 50 L), and
dried at ∼5 Torr and 50 °C 24 h to afford hydroxypyrazole
24b (20.9 kg, 73%) as a white solid that was carried forward
without further purification. The spectral data for this batch was
found to be identical to the laboratory lot described above.
1-(4-Fluorophenyl)-4-isopropyl-5-trifluoromethanesulfo-
nyloxy-1H-pyrazole-3-carboxylic Acid Ethyl Ester (26). To
a solution of hydroxypyrazole 24b (3.2 kg, 10.8 mol) in
dichloromethane (10.4 L)34 was charged triethylamine (1.7 L,
12.2 mol, 1.1 equiv) in one portion, and the solution was cooled
to between -30-20 °C. Trifluoromethanesulfonic anhydride
(3.28 kg, 11.8 mol, 1.1 equiv) was added via addition funnel
1
purity, RRT 12.9 min) as a crystalline white solid. H NMR
(400 MHz, CDCl3) δ 7.50-7.45 (m, 2H), 7.38-7.25 (m, 5H),
7.19-7.12 (m, 2H), 6.80 (d, J ) 4.2, 1H), 6.58 (d, J ) 4.1,
1H), 4.49-4.40 (q, 2H), 3.51-3.43 (m, 1H), 1.48-1.40 (m,
9H). 13C NMR (100 MHz, DMSO-d6) δ 14.17, 22.33, 24.17,
60.90, 115.73, 116.12, 116.26, 127.55, 127.70, 129.05, 129.14,
128.08, 129.98, 136.41, 138.13, 141.87, 161.32, 163.56. MS
(APCI+): 378 [M]. Anal. Calcd for C23H23FN2O2: C, 73.00; H,
6.13; N, 7.40; Obs: C, 73.13; H, 6.10; N, 7.45.
1-(4-Fluorophenyl)-5-formyl-4-isopropyl-1H-pyrazole-3-
carboxylic Acid (31). Laboratory Procedure. To a solution of
styrene30 (789 g, 2.09 mol) in dichloromethane (8.0 L) was
charged methanol (2.0 L), and the solution was cooled with
(35) It should be noted that residual HCl in the Suzuki product likely led
to the formation of the corresponding dimethylacetal in the Bridge
Organics ozonolysis campaign, requiring additional rework.
(34) General attempts toward replacing dichloromethane with alternate
solvents were met with significantly lower yields and new impurities.
154
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

stirring to between -70-60 °C. Ozone36 was charged at a rate
of 10 mL/min < -60 °C. The reaction mixture was monitored
by periodically quenching samples into DCM/DMS for HPLC
analysis. After 1.5 h, there was no more 30 remaining. The
reaction mixture was sparged with nitrogen for 1 h and was
then transferred into a stirred mixture of dimethylsulfide (260
g, 4.18 mol, 2.00 equiv) in dichloromethane (1 L) and stirred
for 1 h at 0 °C. The quenched mixture was concentrated to
dryness at 30 °C and 50 Torr to afford a brown oil (900 g of
crude 11b) that was redissolved in methanol (7.2 L), diluted
with water (360 mL), charged with sodium hydroxide pellets
(432 g, 10.8 mol, 5.17 equiv), and heated to reflux for 4 h. The
resulting solution was concentrated to 2 L at 50 °C and 30 Torr,
diluted with water (12 L), adjusted to pH 6 with concentrated
HCl, and extracted with dichloromethane (3 × 5 L). The organic
extracts were concentrated to 2 L, then charged with toluene
(2 L). The remaining dichloromethane was removed by distil-
lation at 50 °C and <40 Torr, and the resulting toluene solution
was cooled to 0 °C. The solids were collected by filtration,
washed with chilled toluene (2 × 100 mL) and dried for 12 h
at 50 °C and <20 Torr to afford aldehyde 31 (433 g, 75% yield,
>97% HPLC purity, RRT 10.7 min) as a yellow solid. 1H NMR
(400 MHz, CDCl3) δ 9.51 (s, 1H), 7.50-7.43 (m, 2H),
7.25-7.18 (m, 2H), 4.10-3.98 (m, 1H), 1.51-1.40 (m, 6H).
HRMS calcd for C14H13FN2O3: 276.0910, obs: 276.0922.
Flow Ozonolysis Procedure.37 To a solution of styrene 30
(1125 g, 3.0 mol) in dichloromethane (11.4 L) was charged
methanol (2.9 L), and the solution was cooled to -60 °C. The
stock solution was pumped into the top of jacketed (-60 °C)38
continuous flow ozonolysis tube reactor39 at a rate of ca. 15
mL/minute while infusing ozone gas at a rate of 0.4 mol/h. The
rate of ozone infusion was regulated to maintain an excess of
ozone near the bottom of the reactor.40 The reactor output was
transferred into a quench solution containing 4:1 dichlo-
romethane/methanol (2 L) and dimethyl sulfide (252 mL), and
regular TLC analysis (1:1 heptane/EtOAc) of quench solution
aliquots indicated full consumption of 30 (Rf ) 0.3). The quench
solution was concentrated at 30 °C and <50 Torr to 2 L, diluted
with 12 L of water, and extracted with dichloromethane (3 ×
7 L). The organic extracts were concentrated to dryness at 30
°C and <20 Torr to afford 1.63 kg of crude product, which
was identified as the corresponding dimethyl acetal 11c by 1H
NMR. The crude product was treated with methanol (10.5 L)
and water (0.5 L) followed by sodium hydroxide pellets (630
g), and the mixture was heated to reflux for 4 h. The solution
was cooled, concentrated to 6 L at 50 °C and 50 Torr, charged
with water (17.4 L) and was extracted with dichloromethane
(3 × 10 L) to remove impurities. The resulting aqueous phase
was cooled to 0 °C, acidified with conc HCl (1.6 L), and
extracted with dichloromethane (3× 10 L). The organic layer
was concentrated to 8.5 L and was treated with aqueous TFA
(50 wt %, 3.63 L) at 20 °C for 3 h. The resulting mixture was
washed with water (17.5 L), and the aqueous layer was back-
extracted with dichloromethane (3.5 L). The combined organic
layers were washed with aqueous NaCl (10 wt %, 2 × 1.5 L),
dried over magnesium sulfate, filtered, and concentrated to
dryness at 30 °C and >20 Torr to afford a crude solid that was
recrystallized from toluene41 (5.5 L) to afford 493 g (1.78 mol,
60% yield) of 31 as a pale-yellow solid with spectral properties
identical to those above.
Preparation of 1-(4-Fluorophenyl)-5-formyl-4-isopropyl-
1H-pyrazole-3-carboxylic Acid 4-Methylbenzylamide (14).
To a solution of 31 (400 g, 1.45 mol) in dichloromethane (4 L)
was charged N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide
hydrochloride (EDAC-HCl, 411 g, 2.14 mol, 1.5 equiv) at 22
°C, and the mixture was stirred for 25 min. Solid 1-hydroxy-
benzotriazole hydrate (HOBT hydrate, 332.5 g, 2.17 mol, 1.5
equiv), was charged in one portion, and the resulting slurry was
cooled to -5 °C. 4-Methylbenzylamine 13 (240.7 g, 1.99 mol,
1.37 equiv) was added via addition funnel over 30 min at <0
°C. The reaction mixture was warmed to 20 °C for 16 h and
was poured cautiously into sodium bicarbonate (saturated, 2.7
L), taking care to stir well until bubbling subsided. Agitation
was stopped, and the mixture was separated. The organic layer
was washed with additional sodium bicarbonate solution (5 wt
%, 3 L) and sodium chloride (10 wt %, 3 L). The resulting
mixture was charged with acetonitrile (2 L) and was distilled
at 30 °C and 50 Torr to remove dichloromethane. The
acetonitrile solution was distilled at 50 °C and 30 Torr to a
final pot volume of ∼1.6 L, then cooled to 0 °C. The resulting
solids were collected by filtration and dried in a vacuum oven
at 65 °C for 18 h to provide amide 14 (490 g, 89% yield, 97%
1
HPLC purity, RRT 12.1 min) as a tan solid. H NMR (400
MHz, CDCl3) δ 9.99 (s, 1H), 7.38-7.35 (m, 2H), 7.23-7.18
(m, 3H), 7.15-7.11 (m, 3H), 4.54 (d, J ) 5.9 Hz, 2H), 4.16
(sept, J ) 7.0 Hz, 1H), 2.31 (s, 3H), 1.43 (d, J ) 7.2 Hz, 6 H).
19F NMR (376 MHz, CDCl3) δ -111.19. MS (APCI+): 380
[M + H]. Anal. Calcd for C22H22FN3O2: C, 69.64; H, 5.84; N,
11.07. Obs: C, 69.71; H, 5.89; N, 11.17.
3-(tert-Butyl-dimethyl-silanyloxy)-7-[2-(4-fluorophenyl)-
4-isopropyl-5-(4-methyl-benzylcarbamoyl)-2H-pyrazol-3-yl]-
5-oxo-hept-6-enoic Acid Ethyl Ester (33). A mixture of 14
(405 g, 1.07 mol) and ylide 32 (604 g, 1.1 mol, 1.03 eq, >99.5%
ee) in toluene (5 L) was heated to reflux for 36 h until HPLC
analysis indicated full consumption of 14. The pot mixture was
cooled to 20 °C and was concentrated by rotary evaporation to
produce a brown/yellow oil (∼1.1 kg) that was absorbed onto
silica gel (1.1 kg) by concentration from dichloromethane (1.5
L) at 40 °C and 20 Torr. The solids were poured onto a silica
gel pad (1.5 kg),42 and the desired product was eluted with 20%
ethyl acetate/heptane.43 The desired fractions (analyzed by
HPLC to be >97% 33 by area %) were concentrated at 50 °C
and <20 Torr to afford olefin 33 (661 g, 95%, 98.5% HPLC
purity, RRT 16.0 min, >99.5% ee by chiral HPLC) as a pale-
orange oil that was carried forward without further purification.
1H NMR (400 MHz, CDCl3) δ 7.41 (d, J ) 16.4 Hz, 1H),
(36) Benchtop ozone generator. Pacific Ozone, Lab Series L23. Pacific
Ozone, 6160 Egret Court, Benicia, CA 94510, United States. E-mail:
support@pacificozone.com.
(37) Performed at Bridge Organics, 311 W. Washington St., Vicksburg,
MI 49097; http://www.bridgeorganics.com.
(38) Temperature of inner cooling coil.
(41) Cooled to -20 °C to increase recovery of the crystallized product.
(42) Crystallization efforts were cut short due to demands on project
timelines, necessitating silica gel chromatography.
(39) Ozone was introduced 2 in. from the bottom of the tube reactor.
(40) As indicated by a blue solution.
(43) There was 30 L collected in 2.5 L fractions. Fractions with less than
1.5 area % triphenylphosphine oxide contaminant were included.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
155

7.35-7.33 (m, 2H), 7.25-7.23 (m, 3H), 7.18-7.12 (m, 3H),
6.12 (d, J ) 16.3 Hz, 1 H), 4.56 (m, 4H), 4.11 (q, J ) 7.2 Hz,
2H), 3.57 (sept, J ) 7.2 Hz, 1H), 2.74 (m, 2H), 2.48 (m, 1H),
2.32 (s, 3H), 1.45 (d, J ) 7.0 Hz, 6H), 1.30-1.23 (m, 12H),
0.05 (s, 3H), and -0.01 (s, 3H). 19F NMR (376 MHz, CDCl3)
δ -111.61. HRMS calcd for C36H48FN3O5Si: 649.3347, found
649.3340.
MHz, CDCl3) δ 7.36-7.33 (m, 2H), 7.26-7.21 (m, 3H),
7.13-7.08 (m, 3H), 6.46 (d, J ) 16.2 Hz, 1H), 5.64 (d, J )
16.2 Hz, 1H), 4.53 (d, J ) 5.8 Hz, 2H), 4.48-4.46 (m, 1H),
4.23-4.21 (m, 1H), (q, J ) 7.1 Hz, 2H), 3.63 (br s, OH, 1H),
3.49 (sept, J ) 7.0 Hz, 1H), 2.45 (d, J ) 6.0 Hz, 2H), 2.30 (s,
3H), 1.61-1.51 (m, 2H), 1.38 (d, J ) 7.0 Hz, 6H), and 1.26 (t,
J ) 7.0 Hz, 3H). 19F NMR (376 MHz, CDCl3) δ -113.18.
HRMS calcd for C30H36FN3O5 [M]: 537.2639, found 537.2641.
7-[2-(4-Fluorophenyl)-4-isopropyl-5-(4-methyl-benzylcar-
bamoyl)-2H-pyrazol-3-yl]-3,5-dihydroxy-heptanoic Acid Eth-
yl Ester (36). To a stainless steel Parr shaker containing syn-
diol 35 (159 g, 296 mmol) and absolute ethanol (1.5 L) under
nitrogen was charged with Pd/C (10 wt %, 15 g, 50% water).
The reactor was purged with nitrogen followed by pressurization
to 50 psi with hydrogen gas at between 20 and 25 °C, and the
reaction mixture was stirred until hydrogen gas uptake ceased
(4 h). The reaction was purged with nitrogen gas and filtered
through a pad of Celite. The filter cake was rinsed with ethanol
(500 mL), and the filtrates were concentrated to dryness at 50
°C and <20 Torr to afford 151.6 g (95% yield, 98.1% HPLC
purity, RRT 11.1 min, >99% de) of diol 36 as a yellow oil. 1H
NMR (400 MHz, CDCl3) δ 7.37-7.33 (m, 2H), 7.23-7.21
(m, 2H), 7.16-7.10 (m, 4H), 4.52 (d, J ) 6.1 Hz, 2H), 4.15
(q, J ) 7.1 Hz, 2H), 3.74-3.70 (m, 1H), 3.39 (sept, J ) 7.0
Hz, 1H), 2.83-2.78 (m, 1H), 2.68-2.63 (m, 1H), 2.42-2.40
(m, 2H), 2.30 (s, 3H), 1.49-1.41 (m, 3H), 1.39 (d, J ) 7.0 Hz,
6H), and 1.27-1.20 (m, 5H). 19F NMR (376 MHz, CDCl3) δ
-112.47. HRMS calcd for C30H38FN3O5: 539.2795, obs:
539.2801.
7-[2-(4-Fluorophenyl)-4-isopropyl-5-(4-methyl-benzylcar-
bamoyl)-2H-pyrazol-3-yl]-3,5-dihydroxy-heptanoic Acid So-
dium Salt (1). To a solution of ester 36 (204 g 378 mmol) in
absolute ethanol (500 mL) at 15-25 °C was charged aqueous
NaOH (50 wt %, 37.7 g, 471 mmol, 1.25 equiv) in one portion,
and the resulting solution was stirred for 2 h until HPLC analysis
confirmed complete consumption of 36 (less than 0.2% 36 by
HPLC). The reaction mixture was concentrated to remove
ethanol at 50 °C and <20 Torr to produce a light-brown foam
that was redissolved in water (750 mL), warmed to 65 °C, and
treated with a solution of acetic acid (31.1 g, 0.52 mol, 1.38
equiv) in water (100 mL). The resulting mixture was extracted
with ethyl acetate (750 mL, warmed to 65 °C to improve
solubility), and the product-containing organic layer was
separated and set aside. The aqueous layer (pH 6.5) was
readjusted to pH 5 with glacial acetic acid (5 g) and extracted
a second time with ethyl acetate (200 mL). The combined
organic layers were washed with sodium chloride (5%, 200 mL)
and water (200 mL) and concentrated to dryness at 60 °C and
30 Torr to afford a mixture of 37 and 38 as a crude orange
paste that was reconcentrated twice from 2-propanol (2 L) to
dryness (60 °C and 30 Torr) and then dissolved in 2-propanol
(1.12 L) and deionized water (50 mL) at 60 °C. The resulting
solution was held at this temperature and treated dropwise with
NaOH (50 wt %) with stirring over 3 h via transfer pump until
HPLC analysis indicated less than 0.25% lactone 38 remaining
(24.0 g, 300 mmol 50% NaOH solution charged).43 The reaction
mixture was distilled by reflux at ambient pressure to remove
water while backfilling, as necessary, with anhydrous 2-pro-
7-[2-(4-Fluorophenyl)-4-isopropyl-5-(4-methyl-benzylcar-
bamoyl)-2H-pyrazol-3-yl]-3-hydroxy-5-oxo-hept-6-enoic Acid
Ethyl Ester (34). To a solution of TBS-olefin 33 (242 g, 373
mmol, 1.0 equiv) in acetonitrile (2.0 L) held between 0-5 °C
and was added aqueous HF [(CAUTION: use of HF requires
knowledge of handling and disposal information specific to
this material. Do not run this chemistry before reviewing
relevant safety information, including verification of the
proper PPE with a qualified scientist) 48 wt %, 120 mL,
3.31 mol, 8.80 eq] via transfer pump over 30 min. The pot
mixture was warmed to 20 °C and was stirred for 19 h until
HPLC analysis confirmed desilylation was complete (<0.5%
of 33 remaining). The reaction mixture was charged with ethyl
acetate (1 L) and saturated sodium chloride solution (1 L), and
stirred for 10 min. The layers were separated, and the organic
layer was carefully washed with saturated sodium bicarbonate
solution (1.5 L, aqueous layer pH ) 8) followed by sodium
chloride (5 wt %, 1.5 L) and was concentrated to dryness at 60
°C and <20 Torr to afford 186 g (93% yield, 98.5% HPLC
purity, RRT 11.9 min, >99% ee) of ꢀ-hydroxy ketone 34 as an
orange oil that was stored at 0 °C under nitrogen until further
use. 1H NMR (400 MHz, CDCl3) δ 7.45 (d, J ) 16.3 Hz, 1H),
7.36-7.32 (m, 2H), 7.25-7.20 (m, 3H), 7.17-7.12 (m, 3H),
6.10 (d, J ) 16.3 Hz, 1H), 4.56 (d, J ) 6.0 Hz, 2H), 4.47 (pent.,
J ) 7.1 Hz, 1H), 4.19-4.15 (m, 2H), 3.59 (sept, J ) 7.1 Hz,
1H), 3.38 (br s, OH, 1H), 2.74-2.71 (m, 2H), 2.52 (d, J ) 6.3
Hz, 2H), 2.32 (s, 3H), 1.45 (d, J ) 7.0 Hz, 6H), and 1.26 (t, J
) 7.0 Hz, 3H). 19F NMR (376 MHz, CDCl3) δ -111.40. MS
(APCI+): 535 [M]. Anal. Calcd for C30H34FN3O5: C, 67.27; H,
6.40; N, 7.85. Obs: C, 67.10; H, 6.44; N, 7.80.
7-[2-(4-Fluorophenyl)-4-isopropyl-5-(4-methylbenzylcar-
bamoyl)-2H-pyrazol-3-yl]-3-hydroxy-5-oxo-hept-6-enoic Acid
Ethyl Ester (35). To a solution of crude ꢀ-hydroxy ketone 34
(183 g, 341 mmol, 1.0 equiv) in THF (1.8 L) in MeOH (600
mL) at between -78-72 °C was charged diethylmethoxybo-
rane (1.0 M in THF, 444 mL, 444 mmol, 1.3 equiv) in one
portion. The mixture was stirred for 1 h and then treated with
sodium borohydride (15.1 g, 0.4 mol, 1.17 equiv) in four
portions over 45 min while maintaining the pot temperature
below -72 °C. The reaction mixture was stirred for 2 h until
HPLC analysis confirmed full consumption of 34. The reaction
mixture was warmed to -30 °C and was quenched with glacial
acetic acid (68 mL) in one portion, allowing the pot mixture to
warm to 20 °C. The resulting thick solution was charged with
ethyl acetate (1.5 L) and water (1 L), and the phases were
separated. The organic layer was extracted with saturated
sodium bicarbonate (1.5 L) and sodium chloride (5 wt %, 1 L)
and concentrated to dryness at 60 °C and <20 Torr to afford an
oil that was reconcentrated twice from methanol (2 × 500 mL)
to afford 177 g (96% yield, 98.1% HPLC purity, RRT 11.2
min, >99% de) of syn-diol 35 as an orange oil. 1H NMR (400
156
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

panol, until the batch reached 2 wt % water by KF analysis.44
The pot mixture was cooled at a rate of 0.5 °C/min to 15 °C
and was held with stirring for 1 h. The resulting solids were
filtered, washed with 2-propanol (2 × 250 mL), and dried under
vacuum at 40 °C for 22 h to afford 194 g (96%) of 1 as a
granular white solid (96% HPLC purity). The crude product
was recrystallized using a similar procedure from 98:2 IPA/
water to afford 163.3 g (81% yield; >99% HPLC purity, RRT
10.0 min; 99.5% de) of 1 as a crystalline white solid. 1H NMR
(400 MHz, DMSO-d6) δ 8.48 (t, J ) 6.3 Hz, NH), 7.54-7.49
(m, 3H), 7.34-7.30 (m, 1H), 7.15-7.13 (m, 2H), 7.07-7.05
(m, 2H), 4.72 (br s, OH), 4.31 (d, J ) 6.1, 2H), 3.63-3.59 (m,
1H), 3.46-3.43 (m, 1H), 3.21 (sept, J ) 7.0 Hz, 1H), 2.73-2.67
(m, 1H), 2.59-2.53 (m, 1H), 2.22 (s, 3H), 1.94-1.90 (m, 1H),
1.74-1.68 (m, 1H), 1.38-1.29 (m, 3H), 1.26 (d, J ) 7.0 Hz,
6H), and 1.16-1.10 (m, 1H). 19F NMR (376 MHz, DMSO-d6)
δ -113.41. 13C NMR (100 MHz, CD3OD) δ 19.94, 20.10,
21.98, 24.10, 42.99, 43.71, 44.11, 67.03, 68.86, 115.97, 115.99,
126.18, 127.91, 128.11, 129.88, 134.29, 136.50, 142.09, 143.98,
161.15, 163.93, 164.05, 178.91. MS (APCI+): 533 [M - H +
23]. Anal. Calcd for C28H33F1N3NaO5: C, 63.03; H, 6.24; N,
7.86; Na, 4.30. Found: C, 62.62; H, 5.97; N, 7.67; Na, 4.19.
Karl Fischer: 0.14% water.
Acknowledgment
We thank Chemical Development colleagues Dan Belmont,
Tim Curran, Derek Pflum, Tom Nanninga, David Erdman, Sally
Gut, and Randy DeJong and Discovery Chemistry colleagues
Scott Larsen, Gary Bolton, William Park, Mark Bush, Yuntao
Song, and Richard Hutchings for helpful discussions regarding
this project. We thank Russ Linderman, Hayden Thomas, and
Mark Guinn for logistical support, Norm Colbry and Mark
Lovdahl for hydrogenation efforts, Brian Moon for hazards
analysis, Michael Lovdahl, Eric Nord, and Jared VanHaitsma
for analytical assistance, and Michael Stier for outsourcing
support.
(44) The first 95 mol % of hydroxide was added over 10 min with stirring,
and the final amount was titrated in on the basis of HPLC monitoring.
A total of 104 mol % NaOH was required in this campaign to reach
the endpoint, likely due to residual acetic acid.
(45) If necessary, the pot can be distilled until all water is removed at reflux
and then adjusted back to 2 wt % water.
Received for review October 5, 2010.
OP100268E
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
157