Process development of a potent bradykinin 1 antagonist

Article abstract of DOI:10.1021/op8003184

As part of Merck's continued research effort on inflammation and pain, a safe synthesis of an orally bioavailable and CNS penetrant bradykinin 1 antagonist was developed and demonstrated on kilogram scale. The key step included a novel regioselective meta

Full text of DOI:10.1021/op8003184

Organic Process Research & Development 2009, 13, 519–524
Process Development of a Potent Bradykinin 1 Antagonist
Karsten Menzel,*,‡ Fouzia Machrouhi, Matthew Bodenstein, Anthony Alorati,† Cameron Cowden,†,⊥ Andrew W. Gibson,†
Brian Bishop,† Norihiro Ikemoto,⊥ Todd D. Nelson,⊥ Michael H. Kress,§ and Doug E. Frantz|
Merck Research Laboratories, Department of Process Research, Merck & Co., Inc., 466 DeVon Park DriVe, Wayne,
PennsylVania 19087, U.S.A., and Merck Research Laboratories, Department of Process Research, Hertford Road,
Hoddesdon EN11 9BU, United Kingdom
Scheme 1. Retrosynthetic analysis of target molecule (1)
Abstract:
As part of Merck’s continued research effort on inflammation and
pain, a safe synthesis of an orally bioavailable and CNS penetrant
bradykinin 1 antagonist was developed and demonstrated on
kilogram scale. The key step included a novel regioselective
metal-halogen exchange reaction on 1,2-dibromo-5-chloro-3-
fluorobenzene using isopropyl magnesium chloride to install the
1,2,4-oxadiazole ring structure. Suzuki cross-coupling reaction
between a highly functionalized and sterically hindered electrophile
and boronic ester generated the biaryl ring system, which was
converted to the target molecule (1) using standard chemistry. The
safe installation of a 1,2,4-oxadiazole ring proved to be challenging
since the original synthetic route relied on the preparation of a
highly functionalized benzonitrile using potassium cyanide and
resulted in low yields and large amounts of potentially hazardous
waste. Overall, a safe and robust synthesis was developed, which
occurred in eight linear steps with an overall yield of 28%.
Introduction
a short development timeline, it was important to quickly
identify a robust synthetic route which provided multikilogram
quantities.
Kinins play an important role in the pathophysiological
processes of pain and inflammation.1 Bradykinin and kallidin,
two peptide agonists, are known to interact selectively with the
kinin receptor B2, while desArgBradikinin and desArgKallidin
act on kinin receptor B1. Many study results indicate that the
B1 receptor is involved in inflammatory pain and neutrophil
infiltration through activation of a cytokine network.2-4 A highly
selective and potent B1 receptor antagonist could block hype-
ralgesia as shown in animal models. A series of biphenylami-
nocyclopropane carboamide bradykinin B1 receptor antagonists
were discovered and evaluated.5 The structurally related phenyl-
pyridine (1)6was recently targeted for clinical development and
required the preparation of multikilogram quantities. Faced with
Envisioning a convergent route to the target, 1 could be
synthesized from three fragments (A-C) as depicted in Scheme
1. Our process research efforts focused on the development of
an efficient, safe, and robust synthetic route for the synthesis
of fragment A and the coupling reactions between fragments
A-C. The union of A and B could be envisioned by a transition
metal-catalyzed cross-coupling reaction, followed by a typical
amide bond formation with fragment C. The convergent
coupling strategy minimized the consumption of the enantio-
meric enriched fragments B and C and avoided a transition
metal-catalyzed reaction in the last step of the synthesis.
* Author for correspondence. E-mail: karsten_menzel@merck.com.
† Merck Research Laboratories, U.K.
‡ Merck Research Laboratories, Department of Drug Metabolism, Sumney-
town Pike, West Point, PA 19486, U.S.A.
Results and Discussion
Synthesis of Oxadiazole Fragment A. We faced the
challenge to regioselectively install a [1,2,4]-oxadiazole ring
system on a highly functionalized halogenated benzene ring.
After an extensive literature search and bearing the coupling
§ Merck Manufacturing Division, Department of Pharmaceutical Research and
Development, Sumneytown Pike, West Point, PA 19486, U.S.A.
⊥ Merck Research Laboratories, Department of Process Research, 126 East
Lincoln Avenue, Rahway, New Jersey 07654, U.S.A.
| Department of Biochemistry, University of Texas Southwestern Medical
Center, 5323 Harry Hines Blvd., Dallas, TX 75390-9038.
(1) Dray, A.; Perkins, M. N. Trends Neurosci. 1993, 16, 99. Hess, J. F.;
Borokowski, J. A.; Young, G. S.; Strader, C. D.; Ransom, R. W.
Biochem. Biophys. Res. Commun. 1992, 184, 260.
(5) Kuduk, S. D.; et al. J. Med. Chem. 2007, 50, 272.
(6) Wood, M. R.; Schirripa, K. M.; Kim, J. J.; Kuduk, S. D.; Chang, R. K.;
DiMarco, C. N.; DiPardo, R. M.; Wan, B.-L.; Murphy, K. L.; Ransom,
R. W.; Chang, R. S. L.; Holahan, M. A.; Cook, J. J.; Lemaire, W.;
Mosser, S. D.; Bednar, R. A.; Tang, C.; Prueksaritanont, T.; Wallace,
A. A.; Mei, Q.; Yu, J.; Bohn, D. L.; Clayton, F. C.; Adarayn, E. D.;
Sitko, G. R.; Leonard, Y. M.; Freidinger, R. M.; Pettibone, D. J.; Bock,
M. G. Bioorg. Med. Chem. Lett. 2008, 716.
(2) Marceau, F. Immunopharmacology 1995, 30, 1.
(3) Schremmer-Danninger, E.; Offner, A.; Siebeck, M.; Roscher, A. A.
Biochem. Biophys. Res. Commun. 1998, 243, 246.
(4) Decarie, A.; Raymond, P.; Gervais, N.; Couture, R.; Adam, A. Am. J.
Physiol. 1996, 270, 1340.
10.1021/op8003184 CCC: $40.75  2009 American Chemical Society
Published on Web 03/06/2009
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
519

Scheme 2. Regioselective functionalization of 5-bromo-3-
chloro-1-fluorobenzene (2)
efficient and safer with respect to large-scale operations
compared to the route shown in Scheme 3.
Following a modified procedure for the Sandmeyer reac-
tion,10 aniline (3) was converted to the 1,2-dibromobenzene
derivative (7) in the presence of NOBF4 and CuBr2 in a 91%
assay yield.11 The regioselective metal-halogen exchange on
7 was optimized for iPrMgCl (1.2 equiv) in tetrahydrofuran at
-40 °C followed by the addition of dimethylformamide. Upon
aqueous workup at pH 1, benzaldehydes (8 and 9) could be
obtained in 95% assay yield9 as a 97:3 mixture in methyl tert-
butyl ether.12 In agreement with the literature, the bromide at
C-1 of compound 7 was preferentially exchanged due to the
position of the electron-withdrawing fluoride in the ring.
After a solvent switch to 2-propanol for the subsequent step,
benzaldehydes (8 and 9) were treated with an aqueous solution
of hydroxylamine (50%, 1.1 equiv) at 40 °C.13 The expected
benzaldehyde oxime (10) crystallized upon the addition of water
and afforded an off-white crystalline material (98.8 area %) in
75% overall yield from (3). We were pleased to observe that
the undesired regioisomeric oxime, which was formed from 9,
was rejected in that crystallization step. Oxime (10) was
converted to the amidoxime (5) in a two-step process. To this
end, a solution of 10 in dimethylformamide was treated with a
solution of N-chlorosuccinimide (1.05 equiv) in dimethyl-
formamide at 50 °C to afford the imidoyl chloride derivative.13
The resultant dimethylformamide solution of imidoyl chloride
was transferred into a cooled aqueous solution of ammonium
hydroxide (30%, 3.0 equiv) to provide amidoxime (5) in 90%
assay yield.11 Follow-up experiments suggested that, during the
temperature controlled addition of the imidoyl chloride to the
aqueous ammonium hydroxide solution, a succinimide adduct
was formed as the major side product up to 5%.14 We were
pleased to find that this side product could be completely
removed during the isolation of fragment A. After aqueous
workup and solvent switch to 2-propanol, amidoxime (5) was
treated with DMADMA at 20 °C for 90 min to afford fragment
A in 89% assay yield,15,16 which crystallized out of a 2-propanol
solution by the addition of water. Upon a final 2-propanol/water
ratio of 1:1.75, an optimized balance between rejected color
and mother liquor losses was obtained. After filtration and
drying, fragment A was afforded as an off-white solid (45%
isolated yield over six steps; LCAP:17 99.0%, LCWP:17 99.8%;
Kf ) 229 ppm; 1H NMR: traces of IPA; DSC: 83.2 °C, TGA:
0.2% losses).
strategy in mind, we realized that an efficient synthesis could
be derived by the regioselective functionalization of the
halogenated compound (2, Scheme 2) and the sequential
preparation of the 1,2,4-oxadiazole moiety.
Nitrile Route. First we envisioned that the inexpensive and
readily available aniline (3) could be converted to nitrile (4) by
a Sandmeyer reaction as depicted in Scheme 3. In the presence
of NOBF4, 2-bromo-4-chloro-6-fluoroaniline (3) was converted
to the diazonium ion which resulted in the formation of the
benzonitrile (4) in 45-70% yield upon the addition of CuCN
and Cu(BF4)2 in dichloromethane. After an extractive workup
and concentration, the residue was purified by distillation at
120 °C @ 9 Torr to afford a pale-yellow solid. After dissolving
the solid in butanol, the nucleophilic attack of hydroxylamine
at the nitrile carbon atom led to a 9:1 mixture of amidoxime
(5) and amide (6). All attempts failed to improve the product
ratio, and extensive investigations of the experimental conditions
revealed that the amidoxime/amide ratio strongly depended on
the chemical purity, metal content of the starting nitrile (4), and
the choice of solvent. Finally, the amidoxime (5) was condensed
to the 1,2,4-oxadiazole derivative (A) by using dimethylaceta-
mide dimethylacetal (DMADMA) at 22 °C in 20-30% overall
yield starting from the aniline (3).
Our investigations of this synthetic route disclosed several
major challenges for large-scale production, which included:
(I) the evolution of large quantities of nitrogen, (II) evolution
of HCN under pH 1 reaction conditions, (III) evolution of NOx
off-gases, (IV) low yield of nitrile (4), which strongly depended
on the rate of copper tetrafluoroborate addition, (V) distilla-
tion of the nitrile on scale was not practical, and (VI)
amidoxime/amide ratio was highly dependent on the chemical
purity and residual metal level of the nitrile (4) and was in favor
of the undesired amide (6) on multigram scale. In this light an
alternative and less hazardous route was developed for scale.
Regioselective Halogen-Metal Exchange Route. A major
challenge in the synthesis of A is the regioselective synthesis
of the tetrasubstituted benzene ring. Instead of developing an
electrophilic aromatic substitution, an alternative approach was
explored: a previously reported regioselective halogen-metal
exchange of 1,2-dibromobenzene derivatives7-9 was investigated
by trapping the metalated benzene derivative with a carbonyl
compound such as dimethylformamide as outlined in Scheme
4. This synthetic route was evaluated and found to be more
(10) de Lang, R.-J.; van Hooijdonk, M. J. C. M.; Brandsma, L.; Kramer,
H.; Seinen, W. Tetrahedron 1998, 54, 2953.
(11) As determined by HPLC versus a standard isolated, purified, and
1
characterized by H- and 13C NMR.
(12) Interestingly, in contrast to Burton’s observation, in which a rear-
rangement reaction of a lithiated anion in the 3-position of 1,2,4-
trichlorobenzene occured after trapping and aging the reaction mixture
in the presence of DMF, a similar observation with the Grignard
reagent was not observed. Burton, A. J.; Cardwell, K. S.; Fuchter,
M. J.; Lindvall, M. K.; Patel, R.; Packham, T. W.; Prodger, J. C.;
Schilling, M. B.; Walker, M. D. Tetrahedron Lett. 2003, 5653.
(13) Liu, K.-C.; Shelton, B. R.; Howe, R. K. J. Org. Chem. 1980, 3916.
(14) Succinimide side product of 10 was proposed by mass spectrometric
analysis and is believed to be 1-((2-bromo-4-chloro-6-fluoro-phenyl)-
[(E)-hydroxyimino]methyl)pyrrolidine-2,5-dione.
(7) Menzel, K.; Mills, P. M.; Dimichele, L.; Frantz, D. E.; Nelson, T. D.;
Kress, M. H. Synlett 2006, 1948.
(8) Dimichele, L.; Menzel, K.; Mills, P.; Frantz, D.; Nelson, T. Magn.
Reson. Chem. 2006, 44, 1041.
(15) Kocevar, M.; Stanovnik, B.; Tisler, M. J. Heterocycl.Chem. 1982,
1397.
(9) Menzel, K.; Mills, P. M.; Frantz, D. E.; Nelson, T. D.; Kress, M. H.
Tetrahedron Lett. 2008, 49, 415.
(16) Botta, A. Justus Liebigs Ann. Chem. 1978, 306.
(17) LCAP: HPLC area percent; LCWP: HPLC weight percent.
520
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 3. Original synthetic route for the preparation of fragment A
Scheme 4. Final synthetic route for the preparation of fragment A via regioselective halogen-metal exchange
Scheme 5. Cross-coupling reaction between fragment A and B
Cross-Coupling Reaction between Fragments A and B.
The synthesis of the chiral pyridine fragment B has been
recently reported18,19 and was available on multikilogram
quantities. As outlined in Scheme 5, our strategy involved the
cross-coupling between fragments A and B. A robust palladium-
mediated cross-coupling reaction to form the biaryl system had
to be developed. We explored a variety of reaction conditions
for the conversion of fragment B into a nucleophilic coupling
partner, including a potassium trifluoroboronate salt and sub-
sequent cross-coupling reaction under standard reaction condi-
tions.20 However, due to the liberation of HF which caused
corrosion of stainless steel and glassware, this method was not
further pursued, although it provided a reliable process for the
formation of a single boron species. A robust process suitable
(18) Mcwilliams, J. C.; Allwein, S. P.; Nelson, T. D.; O’Shea, Paul; Shultz,
C. S. WO 2006/052514 A1; Chem. Abstr. 2006, 144, 488406.
(19) Allwein, S. P.; McWilliams, J. C.; Secord, E. A.; Mowrey, D. R.;
Nelson, T. D.; Kress, M. H. Tetrahedron Lett. 2006, 6409.
(20) Molander, G. A.; Biolatto, B. J. Org. Chem. 2003, 4302.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
521

Scheme 6. API formation
for scale-up was then discovered by converting the pyridine
into a boronic ester via a palladium-mediated cross-coupling
reaction using bis-pinacol diboron (11).21 The catalyst, ligand,
additives, solvents, and temperature had been optimized to
support a one-pot reaction between fragments A and B. The
initial step required 3 equiv of anhydrous potassium acetate and
3 mol % of Pd(dppf) in toluene at 85 °C to form the boronic
ester (12) quantitatively (based on assay yield) after 12 h. The
cross-coupling reaction was completed by the addition of
aqueous potassium phosphate (1 M; 2 equiv) and the oxadiazole
fragment A (1.05 equiv) in toluene at 90 °C. After 16 h, an
aqueous workup provided the crude biaryl derivative as a
toluene stream in 87% assay yield. The toluene phase was
passed through solka floc to remove black precipitation and was
treated with 50 wt % Ecosorb C-941 for 15 h at room
temperature to reduce color. After filtration, the crystallization
of 13 was initiated by the addition of heptane. A final 2:1
mixture of toluene to heptane minimized the losses to the mother
liquor and afforded 13 in 78% isolated yield.
Formation of Penultimate Compound and API. Although
the hydrolysis of acetamide 13 seemed to be straightforward,
improvements in yield and impurity profile were highly
desirable, since the original procedure provided 14 in low purity
and 75% assay yield using 2 equiv of aqueous 1 M hydrochloric
acid in tetrahydrofuran. In our efforts to identify suitable reaction
conditions, 10 equiv of aqueous 2 M hydrochloric acid
converted acetamide (13) in 9 h at 75 °C to the free amine (14)
as depicted in Scheme 6 in high purity and yield. After aqueous
workup, the amine was isolated in 85% yield as a toluene
stream. Under standard amide bond-formation conditions, target
molecule (1) was synthesized by suspending 1.5 equiv of
1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride
in the toluene stream of amine 14 at room temperature. This
suspension was charged with a solution of hydroxyl benzo-
triazole (0.25 equiv) and (R)-hydroxy acid C22 (1.0 equiv) in
tetrahydrofuran. It was necessary to follow this order of addition
to avoid a self-coupling process of (R)-hydroxy acid (C), which
would result in the formation of dimers and oligomers that could
ultimately couple with the amine to form side products. An
aqueous workup removed the coupling reagents, and the solvent
was switched to toluene. The product crystallized from a
toluene/heptane mixture of 1:1.5, which minimized the losses
to the mother liquor to 3.5%. The final product was filtered
and dried at room temperature, which afforded 2.98 kg of 1 as
a white solid in an 89% yield.
Conclusion
The bradykinin antagonist 1 was synthesized in multikilo-
gram quantities by a convergent route from readily available
starting materials. The synthesis of fragment A was achieved
by a regioselective halogen-metal exchange reaction of 1,2-
dibromobenzene, 7. The subsequent coupling sequence involved
three steps with two isolations and 65% overall yield. Upon
scale-up the synthesis was robust, safe, and efficient in produc-
ing 1 in high purity that could be used for clinical studies.
Experimental Section
Reactions were performed under N2 atmosphere and moni-
tored by HPLC. NMR spectra were obtained at 25 °C in CDCl3,
and the field strengths for the various nuclei were as follows:
1H (700 and 300 MHz), 13C (176 and 75 MHz). Achiral
chromatography was conducted using a YMC-Pack Pro C18
HPLC column (4.6 mm ID × 250 mm, 5 µm). The mobile
phase consisted of A: 1% perchloric acid in deionized water
and B: acetonitrile. The analytes were eluted using a linear
gradient from 10 to 90% B over 25 min at a flow rate of 1
mL/min at room temperature. The analytes were observed by
UV absorption at 220 nm. Chiral chromatography was per-
formed using a Chiralcel OD-H column (4.6 mm ID × 250
mm, 5 µm). The mobile phase consisted of A: isopropanol and
B: hexane. The analyte was eluted using a linear gradient from
5-30% A over 30 min at a flow rate of 0.75 mL/min at room
temperature. The analytes were observed by UV absorption at
220 nm.
1-(2-Bromo-4-chloro-6-fluorophenyl)-N-hydroxymetha-
nimine (10) via 1,2-Dibromo-5-chloro-3-fluorobenzene23 (7)
and 2-Bromo-4-chloro-6-fluorobenzaldehyde (8). 2-Bromo-
4-chloro-6-fluoro aniline (14.5 kg, 64.6 mol) was dissolved in
acetonitrile (102.6 kg) at 20 °C. Nitrosonium tetrafluoroborate
(9.81 kg, 84.0 mol) was added to the reactor in four equal
portions at 0 °C, so that the temperature was maintained at 0
°C before the next portion was added. After the fourth addition,
the reaction mixture was aged for 0.5 h at 0 °C. Acetonitrile
(64.8 kg) and copper (II) bromide (15.9 kg, 71.2 mol) were
combined and stirred for 20 min at 20 °C before the solution
was transferred to the 2-bromo-4-chloro-6-fluoro aniline con-
(21) Selected references for the metal-mediated couplings of pyridine-
containing structures: Hartung, C. G.; Fecher, A.; Chapell, B.;
Snieckus, V. Org. Lett. 2003, 1899. Spivey, A. C.; Zhu, F.; Mitchell,
M. B.; Davey, S. G.; Jarvest, R. L. J. Org. Chem. 2003, 7379.
(22) The acid C was commercially available on kilogram scale from
Lonza. Shaw, N. M.; Naughton, A.; Robins, K.; Tinschert, A.; Schmid,
E.; Hischier, M.-L.; Venetz, V.; Werlen, J.; Zimmermann, T.; Brieden,
W.; Riedmatten, P.; de Roduit, J.-P.; Zimmermann, B.; Neumuller,
R. Org. Process Res. DeV. 2002, 6, 497.
(23) 1-Chloro-3,4-dibromo-5-fluorobenzene is commercially available from
Acros.
522
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

taining reaction mixture so that the batch temperature was
maintained below 5 °C. After completed addition the reaction
mixture was aged for one hour at 0 °C, before the reaction
mixture was warmed to 20 °C and aged for another 4 h. Water
(87 kg) and heptane (49.5 kg) were added sequentially to the
slurry at 20 °C. The organic phase was collected, and the
aqueous phase was extracted with heptane (49.5 kg). After
separation of the aqueous phase, the combined organic phases
were washed with aqueous 2 M hydrochloric acid solution (43.5
kg). The organic phase was concentrated by distillation to an
approximated volume of 44 L (∼16.0 kg of 7, 91% assay yield;
min). 1H NMR (CDCl3, 300 MHz): δ 11.97 (s, 1 H), 8.14 (s,
1 H), 7.76 (s, 1 H), 7.62 (d, J ) 10.3 Hz, 1 H); 13C NMR
(CDCl3, 75 MHz): δ 159.7 (d, J ) 255 Hz), 143.1, 134.8, 128.5,
123.8, 119.9 (d, J ) 15 Hz), 116.2 (d, J ) 23 Hz).
3-(2-Bromo-4-chloro-6-fluorophenyl)-5-methyl-1,2,4-oxa-
diazole (A) via 2-Bromo-4-chloro-6-fluoro-N′-hydroxyben-
zenecarboximidamide (5). The reaction mixture of 10 (8.5 kg,
33.7 mol) in N,N-dimethylformamide (16.0 L) was warmed to
50 °C. A solution of 4.72 kg of N-chlorosuccinimide (35.4 mol)
in 16.1 L of N,N-dimethylformamide was added slowly to the
oxime solution by maintaining the batch temperature between
50 and 60 °C. Upon completed addition, the reaction was
monitored by HPLC and provided a typical assay yield of 95%
of the imidoyl chloride. The reaction mixture was cooled to
3-5 °C and slowly transferred into 4.1 kg of an ammonium
hydroxide solution (2.0 equivalent; 67.3 mol) cooled to 0 °C.
The rate of addition was adjusted to maintain the batch
temperature below +10 °C. After completed addition the batch
temperature was maintained between 0 and 10 °C, and the
reaction progress was monitored by HPLC providing an assay
yield of 90%. Ethyl acetate (15 L) was added to the cooled
reaction mixture, followed by 25 kg of 15 wt % brine. The
reaction mixture was agitated vigorously for 10 min before the
phases were allowed to settle, and the aqueous phase was
separated, and the organic phases were collected. The aqueous
layer was extracted with ethyl acetate (5 L × 2). A solvent
switch of the combined organic layers to 2-propanol (50 L)
was performed, and then the volume was reduced to 200 mg/
mL. DMADMA (8.1 kg, 60.6 mol) was added to the 2-propanol
reaction mixture, and within 15 min the reaction was deemed
completed via HPLC analysis. An addition of 10 wt %/wt Darco
60 was made, and the reaction stirred for one hour before being
filtered through a filter pot containing solka floc. The filter pot
was washed with 5 L of IPA, and the combined organic layers
were charged into a 100-L round-bottom flask. A mixture of
water/IPA, 1:1, was added dropwise to the reaction stream to
start crystallization. After reaching a ratio of water/organic
solvents of 2:1 the white solid was collected by filtration and
dried in a tray dryer under reduced pressure at 66 °C overnight.
Mp: 81.5 °C, HPLC retention time: 15.3 min. 1H NMR (CDCl3,
300 MHz): δ 7.55 (s, 1 H), 7.21-7.27 (m, 1 H), 2.71 (s, 3 H);
13C NMR (CDCl3, 75 MHz): δ 177.4, 163.4, 161.0 (d, J )
255 Hz), 138.2 (d, J ) 15 Hz), 129.3 (d, J ) 3.8 Hz), 124.7
(d, J ) 3.8 Hz), 117.2, 116.4 (d, J ) 15 Hz), 12.7.
1
97.5 LC area %, HPLC retention time: 20.3 min). H NMR
(CDCl3, 700 MHz): δ 7.44 (dd, J ) 2.38, 1.81 Hz, 1H); 7.10
(dd, J ) 7.91, 2.40 Hz, 1H); 13C NMR (CDCl3, 176 MHz): δ
159.5 (d, J ) 228.9 Hz, 1C), 134.4, 128.9, 126.6, 115.9, 111.8.
The heptane solution of 7 (44 L, 16.0 kg, 55.5 mol) was
diluted with tetrahydrofuran (71.0 kg). After cooling the reaction
mixture to -45 °C, isopropylmagnesium chloride solution (31.8
kg; 61.0 mol, 19.8% in tetrahydrofuran) was added over one
hour to the reaction mixture so that the temperature was
maintained between -45 to -40 °C. The batch was aged for
0.5 h at -40 °C before dimethylformamide (20.3 kg, 278 mol)
was added to the reaction mixture over 0.5 h, while the batch
temperature was maintained between -45 and -20 °C. The
reaction mixture was aged for 0.25 h at -20 °C. Deionized
water (187 kg), aqueous concentrated hydrochloric acid (21.1
kg), and methyl tert-butylether (106.6 kg) were combined. The
mixture was cooled to 5 °C before the aldehyde-containing
reaction mixture was added. The batch temperature was
maintained below 5 °C during the addition and upon the
completion was warmed to 20 °C. The organic phase was
separated, and the aqueous phase was extracted two more times
with methyl tert-butylether (106.6 kg) each. All combined
organic phases were washed sequentially with aqueous 1 M
hydrochloric acid (51.2 kg), aqueous 1 M sodium hydrogen
carbonate (51.8 kg), and water (50 kg). After the last aqueous
phase was separated, the organic phase was concentrated by
distillation at reduced pressure to maintain the batch temperature
below 40 °C. The product was obtained in 95% assay yield
and 93 LC area % (HPLC retention time: 14.3 min). 1H NMR
(CDCl3, 300 MHz): δ 9.98 (s, 1 H), 7.53 (s, 1 H), 7.18-7.27
(m, 1 H); 13C NMR (CDCl3, 75 MHz): δ 197.2, 159.7 (d, J )
255 Hz), 138.2, 130.3 (d, J ) 7.5 Hz), 125.8, 123.5, 117.3 (d,
J ) 30 Hz).
A solution of 2-bromo-4-chloro-6-fluorobenzaldehyde (8)
(13.2 kg, 55.5 mol) in methyl tert-butyl ether was charged with
2-propanol (79 kg) and concentrated under reduced pressure
and at a batch temperature of below 40 °C. A second 2-propanol
charge (48 kg) was added, so that the batch volume was around
96 L. The reaction mixture was charged with 50% aqueous
hydroxylamine (4.0 kg) and warmed to 40 °C. After 2 h the
batch was charged with water (48 kg) over a period of one
hour and the slurry was aged for one hour at 20 °C. The solid
was collected by filtration. The wet cake was washed with a
mixture of water and 2-propanol (1:1.5; 17 kg, 26 kg,
respectively) and dried in a tray dryer (40 °C, 50 Torr) to afford
crude 2-bromo-4-chloro-6-fluorobenzaldehyde oxime (10.6 kg,
42.0 mol; 75%; 98.8 LC area %, HPLC retention time: 12.2
N-[(1R)-1-{5-[5-Chloro-3-fluoro-2-(5-methyl-1,2,4-oxadia-
zol-3-yl)phenyl]-3-fluoropyridin-2-yl}ethyl]acetamide (13). A
flask was charged sequentially with N-[(1R)-1-(5-bromo-3-
fluoropyridin-2-yl)ethyl]acetamide (B) (1.5 kg; 5.75 mol),
4,4,4′,4′,5,5,5′,5′-octamethyl-2,2′-bi-1,3,2-dioxaborolane (1.75
kg; 6.9 mol), PdCl2(dppf)·CH2Cl2 (3 mol %, 290 g; 0.35 mol)
and potassium acetate (1.70 kg; 17.3 mol). After purging the
flask with nitrogen gas for one hour, the reaction mixture was
suspended in toluene (7.5 L) and degassed with nitrogen gas
for 30 min. The reaction mixture was heated up to 85 °C and
aged for 18 h. A sample analysis of the reaction mixture by
HPLC determined the conversion to 96%, and the mixture was
allowed to cool to 20 °C. A degassed solution of 3-(2-bromo-
4-chloro-6-fluorophenyl)-5-methyl-1,2,4-oxadiazole (A) (1.76
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
523

kg, 6.04 mol) in toluene (4.4 L) was added to the reaction
mixture, followed by a degassed aqueous potassium phosphate
solution (2 M, 6 L). The batch was heated to 95 °C for 16 h.
The reaction mixture was monitored by HPLC, and upon
completion of the reaction, the mixture was cooled to 50 °C
before water (15 L) was added. The homogeneous reaction
mixture was cooled to 20 °C prior to allowing the phases to
separate and to collecting the aqueous. The organic phase was
washed with aqueous 0.1 M hydrochloric acid (15 L), separated
from the aqueous phase, and charged with Ecosorb C-941 (0.75
kg). The batch was aged for 16 h at 20 °C, before the activated
carbon was collected on a filter pot filled with solka floc. The
filter pot was washed with 5 L of toluene before the combined
organic streams were concentrated to 600 mg/mL at 40 °C under
reduced pressure. At atmospheric pressure and 48 °C, heptane
(8 L) was added slowly to the reaction mixture to start
crystallization. The batch was aged for one hour and allowed
to cool to 20 °C before 72 L of heptane was added to reduce
mother liquor losses and was finally aged for 16 h. After cooling
the slurry to 0 °C, the solids were collected by filtration on a
filter pot. The cake was washed with 2 L of a toluene/heptane
mixture (1:9) and dried under nitrogen for 18 h to afford 1.69
kg as a white solid (4.31 mol; 97.5 LCWP,17 HPLC retention
time: 10.3 min; ee >99% by HPLC (retention time: 6.8 min)).
1H NMR (CDCl3, 300 MHz): δ 8.17 (s, 1 H); 7.25-7.36 (m,
3H), 6.92 (d, J ) 7.38 Hz, 1 H), 5.42-5.51 (m, 1 H), 2.57 (s,
3 H), 2.02 (s, 3 H), 1.47 (d, J ) 6.68 Hz, 3 H); 13C NMR
(CDCl3, 75 MHz): δ 177.1, 169.2, 163.0, 161.0 (d, J ) 255
Hz), 155.4 (d, J ) 263 Hz), 149.1 (d, J ) 15 Hz), 144.1 (d, J
) 3.8 Hz), 140.1, 137.8 (d, J ) 7.5 Hz), 134.3, 126.5 (d, J )
7.5 Hz), 123.7 (d, J ) 15 Hz), 117.0 (d, J ) 30 Hz), 113.9 (d,
J ) 15 Hz), 44.2, 23.4, 21.5, 12.4.
(2R)-3-{[(1R)-1-{5-[5-Chloro-3-fluoro-2-(5-methyl-1,2,4-
oxadiazol-3-yl)phenyl]-3-fluoropyridin-2-yl}ethyl]amino}-
1,1,1-trifluoro-2-methylbut-3-en-2-ol (1) via (1R)-1-{5-[5-
Chloro-3-fluoro-2-(5-methyl-1,2,4-oxadiazol-3-yl)phenyl]-3-
fluoropyridin-2-yl}ethanamine (14). N-[(1R)-1-{5-[5-Chloro-
3-fluoro-2-(5-methyl-1,2,4-oxadiazol-3-yl)phenyl]-3-fluoropyridin-
2-yl}ethyl]acetamide (13) (3.11 kg, 7.92 mol) and 40 L of an
aqueous 2 M hydrochloric acid solution were combined, and
the reaction mixture was warmed to 75 °C, when all solids were
dissolved. After 9 h a sample analyzed by HPLC indicated the
consumption of the starting material. The reaction was cooled
to 20 °C and aged for 16 h. The aqueous stream was extracted
three times with 20 L of a 1:1 mixture of methyl tert-butylether/
heptane. The aqueous stream was diluted with 15 L of toluene
before the mixture was cooled to 8 °C. The pH was adjusted
to 11.3 by the addition of 7 L of an aqueous 10 M sodium
hydroxide solution and 2.9 L of an aqueous 5 M sodium
hydroxide solution. The batch temperature was maintained
below 23 °C by the rate of addition of the sodium hydroxide
solutions. After 0.5 h the phases were allowed to separate and
were collected. The aqueous layer was extracted with an
additional 9 L of toluene. The combined organic layers were
filtered before they were washed with 10 L of an aqueous 15
wt % ammonium chloride solution. The layers were separated,
and the organic layer was concentrated to 85 mg/mL. The
organic stream was charged with 1.28 kg of NuChar. The batch
was aged for 16 h at 20 °C before the activated carbon was
collected on a filter pot filled with Celite 545. The wet cake
was washed two times with 17.5 L of toluene. The combined
organic phases were concentrated again to 85 mg/mL. The
filtration step was repeated one more time, before the batch
was concentrated to 250 mg/mL. HPLC retention time: 4.3 min.
1H NMR (CDCl3, 300 MHz): δ 8.19 (s, 1 H), 7.21-7.29 (m,
3 H), 4.42 (q, J ) 7.0 Hz; 1 H), 2.57 (s, 3 H), 1.87 (s, 2 H),
1.41 (d, J ) 7.0 Hz; 3 H); 13C NMR (CDCl3, 75 MHz): δ 177.0,
163.1, 161.1 (d, J ) 255 Hz), 155.7 (d, J ) 255 Hz), 153.3 (d,
J ) 15 Hz), 144.4 (d, J ) 7.5 Hz), 140.5, 137.7, 133.6, 126.5
(d, J ) 3.8 Hz), 123.2 (d, J ) 7.5 Hz), 116.8 (d, J ) 11.3 Hz),
113.9 (d, J ) 15.0 Hz), 46.4, 23.6, 12.4.
Under nitrogen N-(3-dimethylaminopropyl)-N′-ethylcarbo-
diimide (1.95 kg, 10.2 mol) was charged with 2.38 kg of (1R)-
1-{5-[5-chloro-3-fluoro-2-(5-methyl-1,2,4-oxadiazol-3-yl)phenyl]-
3-fluoropyridin-2-yl}ethanamine (14) as a toluene stream. The
suspension was stirred vigorously at 20 °C. A solution of
1-hydroxybenzotriazole (0.26 kg; 1.69 mol) and (2R)-3,3,3-
trifluoro-2-hydroxy-2-methylpropanoic acid (C) (1.18 kg; 7.45
mol) in tetrahydrofuran (8 kg) was added at a rate that allowed
the batch temperature to remain below 32 °C. The reaction
mixture was agitated for 30 min before the reaction progress
was monitored by HPLC. After 2 h the reaction mixture was
charged with 7.05 L of an aqueous 1 M sodium hydroxide
solution and aged for 0.5 h at 20 °C. The aqueous layer was
collected, and the organic layer was charged with another 4.0
kg of an aqueous 1 M sodium hydroxide solution and agitated
for 5 min. The aqueous layer was collected, and the organic
layer was charged with 7.05 L of an aqueous 1 M hydrochloric
acid solution. The reaction mixture was agitated for 5 min, and
the aqueous layer was collected and discarded. The organic
phase was dried using 7.05 L of 15 wt % brine. The reaction
mixture was agitated for 10 min before the aqueous phase was
collected and discarded. The organic layer was concentrated
prior to the addition of 10 L of toluene and was further
concentrated to 150 mg/mL. The organic stream was passed
through an in-line filter, and the solvent was removed under
reduced pressure at 45 °C to a final concentration of 250 mg/
mL. After the initial addition of 3.6 L of heptane, the product
started to crystallize. The crystallization was completed by the
addition of 8.4 L of heptane. The solid was collected by filtration
on a filter pot and washed with 5 L of a 1:2 mixture of toluene/
heptane. Finally the cake was washed with 2 L of heptane before
the cake was dried under nitrogen for 24 h to afford 2.9 kg as
a white solid (5.9 mol, 87% overall yield). Mp: 121.8 °C, HPLC
retention time: 16.2 min. 1H NMR (CDCl3, 300 MHz): δ 8.22
(s, 1 H), 7.76 (d, J ) 7.14 Hz, 1 H), 7.34-7.37 (m, 2 H),
7.27-7.29 (m, 1 H), 5.44-5.53 (m, 1 H), 4.58 (broad s, 1 H),
2.60 (s, 3 H), 1.61 (s, 3H), 1.49 (d, J ) 6.68 Hz, 3 H); 13C
NMR (CDCl3, 75 MHz): δ 177.7, 167.4, 161.2 (d, J ) 263
Hz), 155.5 (d, J ) 255 Hz), 147.8 (d, J ) 15 Hz), 144.3, 139.9,
137.9 (d, J ) 15 Hz), 134.8, 130.0, 126.5 (d, J ) 7.5 Hz),
124.1 (d, J ) 22.5 Hz), 122.4, 117.2 (d, J ) 22.5 Hz), 113.9
(d, J ) 15 Hz), 74.2 (q, J ) 22.5 Hz), 45.2, 21.0, 20.3, 12.3.
Received for review December 27, 2008.
OP8003184
524
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development