Process development for scale-up of a novel 3,5-substituted thiazolidine-2,4-dione compound as a potent inhibitor for estrogen-related receptor 1

Article abstract of DOI:10.1021/op400325r

The development of a reproducible process for multihundred gram production of (Z)-5-((1-(4-chloro-2-(trifluoromethyl)benzyl)-1H-indazol-5-yl)methylene)-3- ((3R,4R)-3-fluoro-1-methylpiperidin-4-yl)thiazolidine-2,4-dione (26), a potent and selective inhibitor of estrogen-related receptor 1 (ERR1), is described. This multihundred gram synthesis was achieved via magnesium perchlorate- catalyzed regioselective epoxide ring-opening of tert-butyl 7-oxa-3- azabicyclo[4.1.0]heptane-3-carboxylate (9) with thiazolidine-2,4-dione (6, TZD) to form a diastereomeric mixture tert-butyl 4-(2,4-dioxothiazolidin-3-yl)-3- hydroxypiperidine-1-carboxylate (17), of which the 3-hydroxyl group was functionally transformed to 3-fluoro derivative 19 after treatment with Deoxo-Fluor. Chiral separation of 19 provided the desired diastereomer (3R,4R)-21 that was converted to the secondary amine 23 TFA salt. Reductive amination of 23 produced the key intermediate N-methyl 24. Knoevenagel condensation of 24 with 1-(4-chloro-2-(trifluoromethyl)benzyl)-1H-indazole-5- carbaldehyde (5) produced the final product 26 in 10% overall yield (99.7% HPLC area% with ≥99.5% de) after a convergent eight synthetic steps with the only column purification being the chiral HPLC separation of 3R,4R-21 from 3S,4S-22.

Full text of DOI:10.1021/op400325r

Article
pubs.acs.org/OPRD
Process Development for Scale-Up of a Novel 3,5-Substituted
Thiazolidine-2,4-dione Compound as a Potent Inhibitor for Estrogen-
Related Receptor 1
Xun Li,* Ronald K. Russell, Jan Spink, Scott Ballentine, Christopher Teleha, Shawn Branum,
Kenneth Wells, Derek Beauchamp, Raymond Patch, Hui Huang, Mark Player, and William Murray
Janssen Research & Development L.L.C., Welsh and McKean Roads, P.O. Box 776, Spring House, Pennsylvania 19477, United States
ABSTRACT: The development of a reproducible process for multihundred gram production of (Z)-5-((1-(4-chloro-2-
(trifluoromethyl)benzyl)-1H-indazol-5-yl)methylene)-3-((3R,4R)-3-fluoro-1-methylpiperidin-4-yl)thiazolidine-2,4-dione (26), a
potent and selective inhibitor of estrogen-related receptor 1 (ERR1), is described. This multihundred gram synthesis was
achieved via magnesium perchlorate-catalyzed regioselective epoxide ring-opening of tert-butyl 7-oxa-3-azabicyclo[4.1.0]heptane-
3-carboxylate (9) with thiazolidine-2,4-dione (6, TZD) to form a diastereomeric mixture tert-butyl 4-(2,4-dioxothiazolidin-3-yl)-
3-hydroxypiperidine-1-carboxylate (17), of which the 3-hydroxyl group was functionally transformed to 3-fluoro derivative 19
after treatment with Deoxo-Fluor. Chiral separation of 19 provided the desired diastereomer (3R,4R)-21 that was converted to
the secondary amine 23 TFA salt. Reductive amination of 23 produced the key intermediate N-methyl 24. Knoevenagel
condensation of 24 with 1-(4-chloro-2-(trifluoromethyl)benzyl)-1H-indazole-5-carbaldehyde (5) produced the final product 26
in 10% overall yield (99.7% HPLC area% with ≥99.5% de) after a convergent eight synthetic steps with the only column
purification being the chiral HPLC separation of 3R,4R-21 from 3S,4S-22.
discover a nonmicrowave epoxide ring-opening method for
intermediate 9, and (3) to reduce or eliminate the normal
phase chromatographic purification.
INTRODUCTION
■
The Janssen Pharmaceutical R&D medicinal chemists discov-
ered a highly potent and selective inhibitor of estrogen-related
receptor 1 (ERR1), (Z)-5-((1-(4-chloro-2-(trifluoromethyl)-
benzyl)-1H-indazol-5-yl)methylene)-3-((3R,4R)-3-fluoro-1-
methylpiperidin-4-yl)thiazolidine-2,4-dione (26), for the treat-
ment of hyperglycemia in patients with type 2 diabetes mellitus.
In a TR-FRET based assay, compound 26 competitively
displaces a coactivator peptide with an EC50 of 23 nM, whereas
in a cellular two-hybrid luciferase reporter assay, it reduces the
constitutive activity of the receptor with an apparent ED50 of
0.7 μM.1 Initially 50 g of 26 was requested for the rat
tolerability study, which was later followed by an additional
request for 500 g of 26 for monkey tolerability and
cardiovascular toxic studies. The original discovery preparation
of 26, as shown in Scheme 1, was an eight-step synthesis with
overall yield of 4.7%. Some improvement areas for scale-up of
26 were noted: for example, (1) a selective N1-alkylation
process to produce the key intermediate aldehyde 5; (2)
nonmicrowave epoxide ring-opening of compound 9 by
thiazolidine-2,4-dione (6, TZD); (3) improvement of the
isomer mixture of 17/18 to >1:1; (4) removal of the silica gel
column purifications for intermediates 19, 12, 14, 16 and the
final product 26; and (5) remove or move the formaldehyde N-
methylation step. Herein, we report our results for the initial
50-g synthesis of 26 as well as the improved scalable process for
multihundred grams of 26.
The direct alkylation of commercially available 1H-indazole-
5-carbaldehyde (1) with 4-chloro-2-(trifluoromethyl)benzyl
bromide (3), as reported2 resulted in a N1/N2 mixture of 5/
5d (∼1:1). The desired N1 product 5 was very difficult to
separate from the N2 side product 5d by column chromatog-
raphy (Scheme 2). The pure N1 alkylated 5 was alternatively
prepared by a modified three-step regioselective method. In
practice, bromination of 1 using NBS in MeCN afforded 82%
isolated yield of 3-bromo compound 2 with ≥95% chemical
purity.3 Compound 2 was contaminated with a small amount
(∼3−5%) of 1 due to the fast formation of solid 2 that trapped
unreacted 1. Selective N1 alkylation of 2 with 3 in the presence
of K2CO3 (1.2 equiv) resulted in a crude material that
contained ∼88−91% of the desired N1 product 4 and ∼6−8%
of the undesired N2 side product 4a, which was completely
removed by crystallization in EtOAc/heptanes and afforded
pure 4 (>99%, HPLC area%).
The hydrodebromination of 4 using 5−20% Pd/C as the
catalyst resulted in a mixture of 5/5a/5b/5c, and the ratio of 5/
5a/5b/5c changed notably under different reaction conditions.
It was very important to identify a set of conditions that would
afford the highest yield of 5 vs the three side products 5a/5b/
5c, especially deschloro-5b. The presence of 5b in 5 would
originate the side product 27 (Figure 1) after going through the
necessary steps in Scheme 3 or Scheme 5, and the side product
27 is also difficult to remove from the final compound 26. A
number of variables (such as catalyst, reaction time, temper-
RESULTS AND DISCUSSION
■
By working together with medicinal chemistry, our goals for the
50-g campaign of 26 were: (1) to quickly develop a N1-selective
and scalable process for intermediate aldehyde 5, (2) to
Received: November 14, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Original discovery route
Scheme 2. Preparation of compound 5
ature, hydrogen pressure, etc.) that would impact the ratio 5/
5a/5b/5c were investigated extensively (Table 1). The best
reaction conditions for this hydrodebromination was found to
be 5% Pd/C under low hydrogen pressure (5 psi) in the
presence of DIPEA (3.1 equiv) with the mixed solvents of
EtOAc and EtOH (1:1, vol/vol) at 20 °C for 4 h. The water
content in Pd/C catalyst was identified as the most important
factor that governs the ratio of 5/5a/5b/5c with a 5% Pd/C
catalyst containing 35.6% of H2O producing the most
acceptable mixture (typically, in a ratio of ∼90%/∼1%/∼9%/
B
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
with NaHSO3 to remove excess m-CPBA and other peroxide
byproducts, followed by a 10% NaOH wash to remove 2-
chlorobenzoic acid. Addition of 10% excess m-CPBA or
extending the reaction time to 26 h did not push the reaction
to 100% conversion. Epoxide 9 was used without further
purification in next step after drying, since the presence of small
amount of 8 in 9 did not impact epoxide-ring-opening with
TZD derivative 7. Other alternative reagents/conditions for
stereoselective epoxidation of 8 were not investigated.5e
In order to develop a more regioselective and scalable
epoxide-opening condition of 9 to replace the original
microwave promoted conditions, a number of Lewis acid-
catalyzed conditions were studied (Table 2). In general, all
Lewis acid-catalyzed conditions produced the desired 3-OH 10
preferably over the undesired 4-OH 11. Mg(ClO4)2 in DMF
produced the best results and afforded a mixture of 10/11 in a
ratio of 3.2−4.0/1. The solvent DMF was optimum as there
was only a small amount (≤3 HPLC area%) of products 10/11
when the reaction was conducted in either refluxing MeCN,
DCE, dioxane, isopropyl acetate, or MeTHF. Other regiose-
lective epoxide ring-opening conditions6 were not investigated.
After the compatibility of Mg(ClO4)2 in DMF as well as the
thermal stability of the mixture of epoxide 9 and Mg(ClO4)2 in
DMF were checked by TGA that indicated no unexpected
exothermal decompositions. The conditions of Mg(ClO4)2 in
DMF were chosen for scale-up on the strength of the short
reaction time, high ratio of product/regioisomer, and easy
workup. The DMF was removed from the 10/11 mixture by
diluting with water that resulted in >100% mass recovered of
crude 10/11.
Figure 1.
∼1%) as monitored by HPLC and LC/MS. Compound 5 was
isolated in 77% yield with 95−97% of 5 (plus 3−5% of 5b)
after crystallization of the mixture in EtOAc/heptanes that
completely removed 5a and 5c (<0.05% as determined by
HPLC).
Knoevenagel condensation4 of aldehyde 5 with TZD 6, as
shown in Scheme 3, afforded 7 in ≥97% (HPLC area%)
chemical purity, after the resulting crude mixture was slurried in
hot AcOH to remove some of the deschloro aldehyde side
product 5b. This condensation not only afforded a nearly
quantitative yield of the intermediate 7 but also strategically
introduced a strong chromophore at an early synthetic stage
that allowed HPLC and LC/MS to be used for monitoring the
progress of the remaining reactions. This advantage was a trade-
off for the early use of the expensive intermediate 5.
Before proceeding with the scale-up, the compatibility of m-
chloroperoxybenzoic acid (≤77% w/w, m-CPBA) with 8 in
DCM, as well as the thermal stability of epoxide 9 were studied
by RC1 and/or TGA. The results of both safety studies
indicated that this m-CPBA oxidation5 of 8 is a safe reaction
under the given conditions. Treatment of olefin 8 with m-
CPBA from 0 to 20 °C for 20 h resulted in a near quantitative
mass recovery of epoxide 9 that was in 93−95% of epoxide 9
along with 5−3% of 8 and a 0.4−2% (GC area%) of an
impurity in 8 (Scheme 3). The reaction mixture was worked up
Since it was difficult to separate 10 from 11 by normal-phase
column chromatography, the mixture was used “as is” for the
conversion of hydroxyl groups to the corresponding fluoro
Scheme 3. Fifty-gram campaign summary
C
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 1. Selected results of hydrodebromination of the compound 4
entry
catalyst
time (h)
H2O
4
5
5a/5b/5c (HPLC%)
comments
1
Pt/alumina
Pd(dba)2
Pd(dppf)2Cl2
5% Pd/C
5% Pd/C
5% Pd/C
5% Pd/C
5% Pd/C
5% Pd/C
5% Pd/C
5% Pd/C
5
5
5
5
2
2
2
4
4
4
4
none
none
none
none
5 μL
100
10.0
100
0.0
0.0
10.0
0.0/0.0/0.0
0.0/0.0/5.0
0.0/0.0/0.0
12.0/2.0/0.0
0.4/3.7/0.0
0.1/3.1/0.0
0.1/2.5/0.0
11/24/5.3
no rxn (0.25 g of 4)
2
3
0.0
no rxn
4
86.0
Aldrich
5
45.9
63.8
74.4
0.0
50.0
Aldrich
6
10 μL
25 μL
none
none
35.6%
35.6%
33.0
Aldrich
7
23.0
Aldrich
8
59.0
AMC (2 g of 4)
9
0.0
54.0
13/26/5.0
Alfa Aesar (2 g of 4)
Alfa Aesar (2 g of 4)
Alfa Aesar (66 g of 4 × 3)
10
11
0.0
92.5
1.9/5.1/0.5
1.0/8−10/≤1.0
0.0
88.0−91.0
Table 2. Results of Lewis acids-catalyzed epoxide 9 ring-opening with the compound 7
entry
Lewis acid
solvent
temp. (°C)
time (h)
10:11
comments
1
2
3
4
5
6
7
8
9
Mg(ClO4)2
BF3·Et2O
Ba(ClO4)2
Ti(i-PrO)4
CoCl2
DMF
DMAc
DMF
DMAc
DMF
DMAc
DMAc
DMF
neat
115
80
2−6
20
20
72
120
6
3.3−4:1
3.2−4:1
2.2−2.7:1
2.4−2.9:1
3.1−3.8:1
3.3−4:1
2.3−2.8:1
3.6−4.7:1
0:0
115
20
90
incomplete
incomplete
MgBr2·Et2O
80
a
MAB
80
5
Yb(TfO)3
85
24
24
incomplete
(CF3)2CHOH
50
7 is insoluble
a
MAB = methylaluminum bis(2,6-di-tert-butyl-4-methylphenoxide).
Scheme 4. Synthesis of the diastereomer 30 from (3S,4S)-22
functionality using Deoxo-Fluor (bis(2-methoxyethyl)-
aminosulfur trifluoride, 1.5 equiv) in DCM,7 which resulted
in ∼96% formation of a 12/13 mixture (∼3.6−4.3/1) plus
≤1.5% of starting 10/11. The crude 12/13 mixture was
crystallized in acetone/heptane (1/2.5, vol/vol) where most of
the 4-fluoro regioisomer 13 was removed in the filtrate, and the
desired 3-fluoro diastereomeric mixture 12 was isolated in 38%
yield with 94−96% chemical purity. If necessary, column
chromatographic purification of the filtrate provided an
additional 28% of the 12/13 mixture of (∼1/1) along with
15% yield of the cis-impurity mixture analogues of 10/11,
indicating either Lewis acid-catalyzed epoxide 9 ring-opening
with 7 to 10/11 mixture or the fluorination of 10/11 using
Deoxo-Fluor not being 100% stereoselectively controlled.7
The 3-fluoro diastereomeric mixture of 12 (187 g) was
separated by chiral HPLC (ChiralCel-OD, 100% EtOH) to
produce 80.5 g (43.1% isolated yield) of (3R,4R)-14 with 96%
chemical purity and 99.3% de plus 76.0 g (40.6% yield) of the
diasteromer (3S,4S)-15 with 95% chemical purity and 96% de.8
The Boc-protecting group of 14 was cleaved using TFA (10
equiv) to afford quantitatively free amine 16 (99.5% de).
Subsequently, reductive N-methylation of 16 with HCHO/
H2O and NaHB(OAc)3 in DCM afforded the target free base
26 in 96% isolated yield (97% HPLC area% with 99.5% de).
Finally, 26 free base was converted to its HCl salt by treating
with etheral HCl to afford 97% isolated yield of 26 HCl salt as
an amorphous solid. In summary, 26 HCl salt was prepared in
10.6% overall yield with 97% chemical purity and 99.5% de over
eight synthetic steps, with only a chiral chromatographic
D
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 5. Five hundred gram campaign summary
Scheme 6. Proposed mechanism for the ring-contracted side product 20
purification. After this initial campaign, we investigated other
routes for the production of 26 on 500-g scale, because the
chemistry depicted in Scheme 3 (1) used a large quantity of the
expensive aldehyde 5 early in the synthesis, (2) the
intermediates 10, 12, 14, 16, and 26 are photosensitive, and
(3) formaldehyde, a known human carcinogen, was used in the
last step to 26.
The review determined that diastereomer (3R,4R)-24
(Scheme 5) would be a key intermediate for a possible more
cost-effective production of 26 with late stage introduction of
the double bond. This key intermediate could be prepared via
an asymmetric synthesis9 or at least (3R,4R)-21 could be
separated from the undesired (3S,4S)-22 isomer, and the N-
methylation step would be done several steps before API
penultimate 26 to minimize formaldehyde incorporation into
API. Since all the efforts made toward a stereoselective
synthesis of (3R,4R)-21 were inconclusive, the redesigned
synthetic route as shown in Scheme 5 was used to produce the
pure diastereomer 21. In order to prove the concept of Scheme
5, the undesired (3S,4S)-22 was first cleaved to 28 as a TFA salt
E
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
in near quantitative yield (Scheme 4), which was converted to
N-methylated intermediate 29 in 87% isolated yield after
reductive amination. Condensation of 29 with aldehyde 5
resulted in a moderate yield of 30 with the integrity of all
stereogenic centers confirmed by NMR studies.
In this next campaign (Scheme 5), the amount of Lewis acid
Mg(ClO4)2 used for epoxide-ring-opening of 9 was reduced
from 0.7 equiv to 0.21 equiv, with the remaining conditions
similar as before to afford a 90−95% of 17/18 mixture with 2−
4% of 9 and 4−6% of 6. The ratio of 17/18 mixture was
typically in 3.8−4/1. Other solvents such as 1,4-dioxane,
isopropyl acetate, methylTHF, hexafluoroisopropanol, and
NMP were investigated but none was better than DMF.
Compound 17 was obtained in 97−99% chemical purity (plus
1−3% of 18) after two crystallizations of the initial crude
reaction mixture in EtOAc/heptanes (1:3).
Furthermore, the amount of Deoxo-Fluor was reduced from
1.35 equiv to 1.1 equiv, leading to a ∼4.3/1 19/20 mixture with
<1% of starting 17. The purity of 19 could be improved to 94−
96% by crystallization of the crude from acetone/heptane (1/
2.5) with 50% recovery of 19. Later, an improved purification
was found by directly triturating the crude 19/20 in EtOAc to
provide >98% purity 19 with a slightly better isolated yield
(55%). The structure of the ring-contracted side product 20
was elucidated by HNMR and confirmed by X-ray diffraction
studies, and a proposed mechanism10 for 20 is given in Scheme
6.
A chiral HPLC method was developed to separate (3R,4R)-
21 from (3S,4S)-22 that was transferred to the Janssen Center
of Preparative Separation Techniques at Beerse, Belgium, which
supplied a total 1333 g (45% recovery yield) of the desired
(3R,4R)-21 in 97−99% chemical purity (HPLC area%) with
99.1−99.6% de (chiral HPLC) from 2951 g of diastereomeric
mixture of 19.
CONCLUSIONS
■
In summary, a reproducible process was developed for large-
scale production of 26, which afforded 800 g of 26 HCl salt was
prepared in 9% overall yield with 99.7% chemical purity and
≥99.5% de over eight synthetic steps, with only one chiral
chromatography required for the separation of (3R,4R)-21
from (3S,4S)-22. The 500-g route is improved as compared to
the initial campaign in that: (1) formaldehyde is now
introduced earlier in the process, (2) the required amount of
costly aldehyde 5 is reduced by six, and (3) the 500-g route
results in much less (≤0.3%) of the undesired E-isomer. This
process has been repeated successfully at the same scale.
EXPERIMENTAL SECTION
■
The purity/impurity ratio of compound 9 was determined on
an Agilent 6890 series GC System with 7639 series injector and
a HP-5 5% phenyl methyl siloxane column (dimensions =
capillary 30.0 m × 320 μm × 0.25 μm nominal), oven
temperature = 50 to 210 °C, ramp = 10 °C/min, detector
heater = 250 °C; H2 flow = 40 mL/min, air flow = 450 mL/
min, N2 flow = 25 mL/min.
The purity/impurity ratios of all compounds were
determined on an Agilent series 1100 system at 254 nm
(UVmax) and with specified columns and conditions as
described below.
A Phenomenex Luna C18 (2) column (4.6 mm ID × 50 mm,
5.0 μm) was used for analyzing compounds 2, 4, 5, and 26 (as
well as 26·HCl salt) at 35 °C with using solvents: A = H2O +
0.1% TFA, B = CH3CN + 0.1%TFA at flow rates of 1.0 mL/
min. The run times for compounds 2, 4, and 5 were 10.0 min
with Gradient: B 15%/0.0 min, B 50%/1.0 min, B 85%/4.0 min,
B 85%/8.0 min, B 15%/10.0 min. The run times for compound
26 were 12.0 min with Gradient: B 15%/0.0 min, B 50%/1.0
min, B 85%/4.0 min, B 85%/10.0 min, B 15%/12.0 min. The
run times for compound 24 were 20 min with solvents: A =
H2O, B = CH3CN and Gradient: B 30%/0.0 min, B 70%/6.0
min, B 90%/12.0 min, B 90%/18.0 min, B 30%/20.0 min.
A Zorbax Eclipse XDB-phenyl column (4.6 mm ID × 50
mm, 3.5 μm) was used for analyzing compounds 17, 19, 21,
and 23 TFA salt at 35 °C with solvents: A = H2O + 0.1% TFA,
B = CH3CN + 0.1%TFA at flow rate of 1.0 mL/min. The run
times for compound 17 were 26.0 min with Gradient: B 6%
/0.0 min, B 20%/20.0 min, B 90%/25.0 min, B 6%/26 min.
The run times for compounds 19, 21, and 23 TFA salt were
20.0 min with Gradient: B 6%/0.0 min, B 30%/8.0 min, B
90%/18.0 min, B 6%/20 min.
An analytical ChiraPack AD column (4.6 mm ID × 250 mm,
3.5 μm) was used to determine the optical purity of compounds
17, 19, and 21 at 25 °C with solvents: A = EtOH + 0.05% DEA,
B = heptanes (A/B = 7/3). The run time was 30.0 min with an
isocratic flow rate of 0.6 mL/min.
An analytical ChiraPack AD-H column (4.6 mm ID × 250
mm, 3.5 μm) was used to determine the optical purity of
compounds 23 and 24 at 25 °C with solvents: A = EtOH +
0.1% DEA, B = heptanes (A/B = 9/1). The run time was 30.0
min with an isocratic flow rate of 1.0 mL/min, whereas for
compound 26 and 26·HCl salt it was with solvents: A = EtOH
+ 0.05% DEA, B = heptanes (A/B = 6/4). The run time was
50.0 min with an isocratic flow rate of 0.6 mL/min.
For this campaign, the amount of TFA was lowered to 5.5
equiv for cleaving the Boc group, which afforded (3R,4R)-23
TFA salt in quantitative yield (ratio of (3R,4R)-23/TFA = 1/1,
confirmed by elemental analysis) with >97% chemical purity
and 99.5% de, which was used in the next step without further
purification.
For the reductive N-methylation, the (3R,4R)-23 TFA salt
was treated in situ with 1.23 equiv of Na2CO3 to give free base
23, which was subsequently treated with HCHO/H2O and
NaHB(OAc)3 (both reduced from 2.0 equiv to 1.35 equiv) to
afford 83−89% isolated yield of (3R,4R)-1-methyl 24 with 98%
chemical purity and diastereomeric purity of 99.2% de.
Knoevenagel condensation of (3R,4R)-24 was performed in
the dark with an equal equivalent of 5 (98.8% wt% with 1.2%
deschloro side product 5b) in the presence of 0.53 equiv of
piperidine in toluene at 110−112 °C over 72−86 h. The
resulting desired 26 free base was obtained in 82−91% isolated
yield with 99.5% wt% and 99.7% de simply after collecting the
solid 26 from the cold reaction mixture and washing with
hexanes. The photoisomerization of (Z)-26 to E-byproduct was
minimized from previously ∼3.0% to ≤0.3% as determined by
both RP-HPLC and chiral HPLC.
Finally, 26 free base in THF was converted to its HCl salt
after treatment with 2.0 equiv of 1.0 M HCl in Et2O to afford
26 HCl salt in 97% isolated yield with 99.7% wt% and 99.7% de
as an amorphous solid.
3-Bromo-1H-indazole (2). 1H-Indazole-5-carbaldehyde
(1) (250.0 g, 1.71 mol) and acetonitrile (5.00 L) were stirred
for 40 min at 20 °C until it became a clear solution. N-
F
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Bromosuccinimide (370.9 g, 2.05 mol) was added portionwise
over 5 min at 20 °C (an exotherm was observed from 20 to 34.9
°C within 15 min). A precipitate formed almost immediately,
and the thick slurry was stirred for 4 h at 20 °C, and the
progress was monitored by HPLC and LC/MS. The resulting
thick, tan mixture was cooled to 10 °C, and the solid was
collected by filtration. The filter cake was washed with MeCN
(500 mL) followed by deionized water (1 L × 2), air-dried
overnight (HPLC indicated the aqueous MeCN filtrate
contained 6−8% of product 2 while the tan solid showed
∼5% of starting 1), and then the resulting damp tan solid was
placed in a vacuum oven at 50−55 °C at 10 mmHg for 2.5 days
until TGA determined a ≤ 0.3% weight decline. There was
obtained 318.0 g (82% isolated yield; 95% of 3-bromo 2 plus
5% of starting 1; HPLC area%) of compound 2 as an off-white
containing 35.6% of water) was introduced. The flask was set at
5 psi at 20 °C for 3 h and the progress of the reaction was
monitored by HPLC and LC/MS (until the starting 4 was
≤1%). The catalyst was removed by filtering through a Celite
545 pad, which was washed with EtOAc (120 mL × 3). The
combined filtrate was concentrated at 66 °C under house
vacuum to give the crude material (96.3 g that contained
DIPEA+Br− salt) which was redissolved in EtOAc (660 mL)
and washed with D.I. H2O (360 mL × 2). The organic phase
was separated and concentrated at 66 °C at 10 mmHg to give
56.1 g (97% isolated yield, 89% of 5, 1.6% of 4, a total 9.4% of
reduced alcohol 5a and des-chloro 5b, and <0.3% of des-chloro
alcohol 5c; HPLC area%) of crude 5 as a clear yellowish oil.
The above reaction was repeated twice in the same scale that
afforded a total 159.9 g of crude 5 after combination. This
crude 5 (159.9 g) was dissolved in EtOAc (220 mL) at 73 °C,
heptane (1.0 L) was added over a 5-min period and stirred at
73 °C for 5 min, gradually cooled to 20 °C over 2 h while
seeded with crystalline solid 5 at 66 to 56 °C. The resulting
white slurry was stirred at 20 °C for an additional one hour and
the solid was collected by filtration, washed with heptane (80
mL × 2), and dried under air-suction for 72 h to afford 127.6 g
(77.6% isolated yield, 95% of 5, 0.5% of 4, 0% of des-chloro
alcohol 5c, and a total 4.5% of reduced alcohol 5a and des-
chloro 5b; HPLC area%) of 5 as an off-white crystalline solid.
Mp = 113−114 °C. 1H NMR (300 MHz, DMSO-d6) δ 5.90 (s,
2 H), 6.78 (d, J = 8.31 Hz, 1 H), 7.66 (dd, J = 1.71, 8.56 Hz, 1
H), 7.83 (d, J = 8.56 Hz, 1 H), 7.87−7.96 (m, 2 H), 8.47 (s, 1
H), 8.51 (s, 1 H), 10.07 (s, 1 H, CHO). 13C NMR (400 MHz,
DMSO-d6) δ 192.09, 142.12, 136.55, 134.22, 132.88, 132.86,
132.93, 130.47, 128.04, 127.39, 126.11, 125.20, 124.66, 123.50,
110.44, 48.25. LC/MS m/z 339.0 (M+).
1
solid. Mp = 115−116 °C. H NMR (300 MHz, DMSO-d6) δ
7.73 (d, J = 8.80 Hz, 1 H), 7.92 (dd, J = 1.26, 8.60 Hz, 1 H),
8.29 (s, 1 H), 10.09 (s, 1 H, CHO), 13.91 (s, 1 H, NH). 13C
NMR (400 MHz, DMSO-d6) δ 192.20, 143.20, 129.99, 127.66,
124.37, 122.58, 122.36, 110.92. LC/MS m/z 256.9 (MH+),
472.0 (2M + Na)+.
3-Bromo-1-(4-chloro-2-(trifluoromethyl)benzyl)-1H-
indazole-5-carbaldehyde (4). 3-Bromo-1H-indazole (2,
150.0 g, 0.65 mol), DMF (1.27 L), and potassium carbonate
(108.9 g, 0.78 mol) were stirred at 20 °C. 1-(Bromomethyl)-4-
chloro-2-(trifluoromethyl)benzene (3) (199.5 g, 0.72 mol) was
added over 22 min (an exotherm was observed from 20 to 37.8
°C after the addition). The reaction was stirred for 30 min. The
progress was monitored by HPLC and LC/MS. The mixture
was poured into H2O/EtOAc (1.48 L/5.84 L) and stirred for 5
min at 20 °C. After phase separation, the organic layer was
washed with 0.5 N HCl (2 × 600 mL) and brine (6 × 800 mL).
The solvent was concentrated at 66 °C under house vacuum
and then to 73 °C at 20 mmHg to afford 291.9 g (108% mass
recovery; 91% of 4 and 8.1% of N2 alkylated 4a; HPLC area%)
of crude 4 as a thick, yellowish oil. This crude 4 was dissolved
in in EtOAc (240 mL) and warmed to 60 °C with stirring.
Warm heptane (4.6 L at 50 °C) was added over 10 min, and
the resulting slightly yellowish, clear solution was stirred at 60
°C for 5 min and then gradually cooled to 20 °C over 3 h while
seeded with crystalline 4 at 53 and 48 °C and stirred one
additional hour at 20 °C. The white solid was collected by
filtration, washed with heptane (250 mL × 2), and dried in a
vacuum oven at 50 °C to afford 203 g (75% isolated yield;
99.7% of 4, HPLC area%) of 4 as a white, fluffy, crystalline
solid. (An additional 6.9 g of an off-white solid was collected
from the filtrate after standing at 20 °C for 72 h, of which
HPLC determined a mixture of 24% N1 product 4 and 76% N2
byproduct 4a.) Mp = 90−91 °C. 1H NMR (300 MHz, DMSO-
d6) δ 5.90 (s, 2 H), 6.96 (d, J = 8.56 Hz, 1 H), 7.70 (dd, J = 2.2,
8.56 Hz, 1 H), 7.88−7.96 (m, 2 H), 8.01 (dd, J = 1.22, 8.80 Hz,
1 H), 8.34 (s, 1 H), 10.11 (s, 1 H, CHO). 13C NMR (400
MHz, DMSO-d6) δ 192.00, 143.42, 133.56, 133.16, 133.01,
131.43, 131.27, 128.03, 126.30, 124.61, 123.34, 123.05, 121.88,
119.15, 111.36, 48.81. LC/MS m/z 417.0 (M+), 419.0 (M+ +
2), 440.9.0 (M + Na)+.
tert-Butyl 7-Oxa-3-azabicyclo[4.1.0]heptane-3-car-
boxylate (9). m-Chloroperoxybenzoic acid (1070.1 g, 4.78
mol, ≤77% w/w) in DCM (5.63 L) was cooled to 0 °C; a
solution of 8 (700.0 g, 3.82 mol) in DCM (5.63 L) was added
as a small stream at 1−4 °C over 3 h. The resulting suspension
was stirred at 0 °C for 30 min, gradually warmed to 10 °C over
a 2-h period, and then stirred at 20 °C for 20 h. The progress of
the reaction was monitored by GC, LC/MS, and TLC
(heptane/EtOAc, 4:1 and I2 chamber). The insoluble solid
(3-chlorobenzoic acid) was removed by filtration, and the
filtrate was stirred with 20% NaHSO3 solution (2.2 L) for 20
min (peroxides in the reaction mixture were monitored both
organic and aqueous phase using starch iodide test strip). After
phase separation, the organic layer was washed with 10%
NaOH solution (2.2 L × 3) and concentrated at 36 °C under
house vacuum and then at 10 mmHg to afford 741.0 g (97%
isolated yield, 95.8% of 9, 2.3% of starting 8, and 1.3% of one
uncharacterized byproduct that was carried over from starting
8; GC area%) of epoxide 9 as a colorless oil, in which
containing traces of starting 8 and an unknown aromatic from
olefin 8 as confirmed by HNMR study. Compound 9 was used
1
in next step without further purification. H NMR (300 MHz,
CDCl3) δ 1.45 (s, 9 H, 3 CH3), 1.86−1.96 (m, 1 H), 1.98−2.14
(m, 1 H), 3.04−3.16 (m, 1 H), 3.16−3.26 (m, 1 H), 3.26−3.31
(m, 1 H), 3.37−3.51 (m, 1 H), 3.59−3.76 (m, 1 H), 3.76−3.98
(m, 1 H). 13C NMR (400 MHz, CDCl3) δ 166.30, 79.70, 58.50,
50.41, 42.62, 37.15, 28.48 (3 C), 24.44. LC/MS m/z 144.1
(MH − t-Bu)+ of 9, 222.1 (M + Na)+, 421.2 (2M + Na)+.
tert-Butyl 4-(2,4-dioxothiazolidin-3-yl)-3-hydroxypi-
peridine-1-carboxylate (17). tert-Butyl 7-oxa-3-
azabicyclo[4.1.0]heptane-3-carboxylate (9, 1110 g, 5.25 mol),
1-(4-Chloro-2-(trifluoromethyl)benzyl)-1H-indazole-5-
carbaldehyde (5). 3-Bromo-1-(4-chloro-2-(trifluoromethyl)-
benzyl)-1H-indazole-5-carbaldehyde (4, 71.4 g, 0.171 mol),
ethyl acetate (285 mL), ethanol (285 mL), and diisopropyl-
ethylamine (92.16 mL, 0.528 mol) were placed in a 2.25-L Parr
hydrogenation bottle and flushed with nitrogen. Palladium
catalyst (2.14 g, 1.00 mmol; 5% on activated carbon that
G
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
DMF (1.81 L), thiazolidine-2,4-dione (6) (500.0 g, 4.27 mol),
and magnesium perchlorate (202.2 g, 0.897 mol) were stirred at
20 °C for 20 min (there was an exotherm to 36.4 °C after the
addition of Mg(ClO4)2). The reaction was gradually warmed to
115 °C over a 2-h period and was stirred at 115 °C for 10 h.
The progress of the reaction was monitored by HPLC and LC/
MS until the ratio of desired 17 to undesired 18 (HPLC area%)
had no further change. The reaction was cooled to 20 °C, and
EtOAc (14.5 L) was added and stirred for 5 min. The organic
phase was washed with D.I. water (3.0 L × 3, 2.0 L × 2) and
dried over MgSO4. After filtration, the organic phase was
concentrated at 53 °C under house vacuum to give a
concentrated solution (∼2.0 L) that was diluted with EtOAc
to the final total volume = 2.70 L. This resulting solution was
warmed to 70 °C with stirring; heptane (3.0 L) was added as a
small stream over 60 min, and the mixture stirred for 5 min at
70 °C and then gradually cooled to 20 °C over 6 h after seeding
with pure 17 (98.5%) at 66 °C. After the resulting suspension
was stirred at 20 °C for 18 h, it was cooled to 14 °C, stirred for
3 h, and further cooled to 10 °C and stirred for an additional 2
h. The solid was collected by filtration, washed with heptane
(300 mL × 2), dried under air-suction for 20 h and then in a
vacuum oven at 50 °C for 20 h to afford 782.0 g (58% isolated
yield, 73.5% of 17, 8.7% of 18, 2.6% of starting 6, 2.0% of
starting 9; HPLC area%) of crude 17 as a yellow powdery solid.
The above obtained crude 17 (782.0 g) was suspended in a
mixed solution of EtOAc/heptane (0.48 L/1.43 L) with stirring
and was heated to 73−76 °C for 30 min. The resulting mixture
was cooled to 20 °C and stirred for 18 h. The solid was
collected by filtration, dried by air-suction for 2 h, and then
placed in a vacuum oven at 50 °C for 20 h to afford 595.0 g
(44% isolated yield, 97% of 17, 1.8% of 18; HPLC area%) of
94.1% of 19, 3.9% of 20, 1.9% of 17; HPLC area%) of
compound 19 as a slightly yellow solid. The filtrate was
concentrated to give 116.0 g of an oily, yellowish solid, which
was recrystallized in EtOAc (116 mL) to afford an additional
81.5 g (12% isolated yield; 98.0% of 19, 2.0% of 20; HPLC area
%) of compound 19 as a yellowish solid. Mp = 151−152 °C.
1H NMR (300 MHz, CDCl3) δ 1.47 (s, 9 H, 3 CH3), 1.75−
1.65 (m, 1 H), 2.38 (ddd, J = 4.40, 4.65, 12.71 Hz, 1 H), 2.62−
2.83 (m, 2 H), 3.97 (s, 2 H), 4.08−4.26 (m, 1 H), 4.28−4.38
(m, 1 H), 4.41−4.66 (m, 1 H), 5.15 (ddd, J = 5.6, 10.27, 50.13
Hz, 1 H). 13C NMR (400 MHz, CDCl3) δ 171.65, 171.36,
154.10, 84.00, 80.70, 57.07, 47.03, 42.57, 33.27, 28.37 (3 C),
26.50. LC/MS m/z 319.4 ((MH)+ not detected), 341.1
(MNa)+, and 219.1 [M − Boc]+. Calcd for C13H19FN2O4S
(MW = 318.36): C, 49.09; H, 6.02; N, 8.80; F, 5.97; S, 10.07.
Found: C, 49.15; H, 6.33; N, 8.78; F, 5.93; S, 10.37.
Chiral Separation of 19 to (3R,4R)-tert-Butyl 4-(2,4-
Dioxothiazolidin-3-yl)-3-fluoropiperidine-1-carboxylate
(21) and (3S,4S)-22. The diastereomeric compound 19 (total
2951.0 g) was separated by using preparative chiral HPLC at
the Center of Preparative Separation Techniques at Beerse,
Belgium, to afford 1333.0 g (45% yield) of (3R,4R)-21 in 95.5−
99.0% chemical purity with 99.1−99.7% de (plus 93.0 g of
(3S,4S)-22 in 95% chemical purity with 97% de). Preparative
chiral HPLC conditions: Chiralpak AD (1000 Å; 20 μm; 2000
g, ID = 5 cm × L = 25 cm) preparative column with Knauer
superpreparative cell (0.5 mm) detector at 254 nm (UVmax) at
30 °C; solvent = 100% MeOH with an isocratic flow rate = 750
1
mL/min. Mp = 110−111 °C. H NMR (300 MHz, CDCl3) δ
1.47 (s, 9 H, 3 CH3), 1.75−1.65 (m, 1 H), 2.38 (ddd, J = 4.40,
4.65, 12.71 Hz, 1 H), 2.62−2.83 (m, 2 H), 3.97 (s, 2 H), 4.08−
4.26 (m, 1 H), 4.43 (ddd, J = 4.40, 4.65, 10.51 Hz, 1 H), 4.41−
4.66 (m, 1 H), 5.15 (dddd, J = 5.38, 5.62, 10.3, 50.1 Hz, 1 H).
13C NMR (400 MHz, CDCl3) δ 171.65, 171.36, 154.10, 84.00,
80.64, 57.08, 46.95, 42.59, 33.28, 28.32 (3 C), 26.56. LC/MS
m/z 319.4 ((MH)+ not detected), 341.1 (MNa)+, and 219.1
[M − Boc]+. Calcd for C13H19FN2O4S (MW = 318.36): C,
49.09; H, 6.02; N, 8.80; F, 5.97; S, 10.07. Found: C, 49.15; H,
6.33; N, 8.78; F, 5.93; S, 10.37.
1
pure 17 as a beige-to-yellow solid. Mp = 143−144 °C. H
NMR (300 MHz, CDCl3) δ 1.46 (s, 9 H, 3 CH3), 1.70−1.60
(m, 1 H), 1.86−2.03 (m, 1 H), 2.33 (ddd, J = 4.65, 4.65, 12.72
Hz, 1 H), 2.45−2.64 (m, 1 H), 2.65−2.85 (m, 1 H), 3.96 (s, 2
H), 4.03−4.16 (m, 2 H), 4.27−4.42 (m, 2 H). 13C NMR (400
MHz, CDCl3) δ 172.20, 171.97, 154.36, 80.39, 65.08 59.39,
49.79, 43.05, 33.31, 28.39 (3 C), 26.86. LC/MS m/z 317.1
(MH)+, 339.1 (MNa)+, and 217.1 [M − Boc]+. Calcd for
C13H20N2O5S (MW = 316.37): C, 49.35; H, 6.37; N, 8.85; S,
10.14. Found: C, 49.45; H, 6.50; N, 8.85; S, 10.13.
(3R,4R)-4-(2,4-Dioxothiazolidin-3-yl)-3-fluoropiperi-
din-1-ium TFA Salt (23). To a solution of the diastereomer
(3R,4R)-21 (500.0 g, 1.57 mol) in DCM (4.0 L) stirred at 20
°C, was added dropwise trifluoroacetic acid (594.0 mL, 7.86
mol) over 20 min. The reaction was stirred for 6 h until HPLC
determined ≤1% of starting 21, and LC/MS detected no more
fragment 23a (MH+ = 263.1). The solvent was removed at 36
°C under house vacuum and then at 66 °C at 10 mmHg to
afford 527.3 g (100% isolated yield; 98% of (3R,4R)-23, HPLC
area%) of (3R,4R)-23 TFA salt as a yellow-to-beige solid, which
was used “as is” in the next step without further purification.
tert-Butyl 4-(2,4-Dioxothiazolidin-3-yl)-3-fluoropiperi-
dine-1-carboxylate (19). To the cold solution of compound
17 (700.0 g, 2.18 mol) in DCM (11.0 L) at 0 °C, was added
dropwise a solution of bis(2-methoxyethylaminosulfur tri-
fluoride) (426.0 mL, 2.31 mol) in DCM (1.0 L) over 1 h.
The resulting mixture was stirred at 0 °C for 2 h and then
stirred at 20 °C for 20 h. The progress of the reaction was
monitored by HPLC and LC/MS. The reaction was cooled to 0
°C with fast agitation, and a solution of 2 N NaOH (2.0 L) was
carefully added (Caution: Addition of the first 500 mL of 2 N
NaOH solution is very exothermic; the internal temperature was
maintained between 0 to 10 °C by adjusting the addition rate).
The resulting suspension was allowed to stir for 1 h at 0 to 10
°C. After phase separation, the organic phase was washed with
saturated NaHCO3 (2.0 L) and D.I. H2O (2.0 L) and then was
dried over Na2SO4. The solvent was concentrated to dryness
under house vacuum to give the crude 19 as an oily material.
This crude 19 was triturated with EtOAc (1.0 L) at 25 °C for
30 min, the solid was collected by filtration, and the filter cake
was washed with heptanes (300 mL × 3) and dried in a vacuum
oven at 50 °C for 20 h to afford 382.0 g (55% isolated yield;
1
Mp = 243−244 °C dec. H NMR (300 MHz, CDCl3) δ 1.98
(br d, J = 13.9, Hz, 1 H), 2.42 (ddd, J = 3.67, 3.91, 13.5 Hz, 1
H), 3.09 (dd, J = 2.45, 13.5 Hz, 1 H), 3.18 (ddd, J = 4.65, 6.36,
11.3 Hz, 1 H), 3.35 (d, J = 12.2 Hz, 1 H), 3.70 (d, J = 14.7 Hz,
1 H), 4.27 (s, 2 H), 4.57 (ddd, J = 3.67, 4.40, 12.0 Hz, 1 H),
5.37 (dddd, J = 4.89, 5.13, 10.5, 49.2 Hz, 1 H) 8.79−9.95 (br s,
2 H). 13C NMR (400 MHz, DMSO-d6) δ 172.44, 172.03,
158.42, 116.99, 82.89, 53.38, 44.35, 41.94, 33.39, 22.81. LC/MS
m/z 219.1 ((MH)+ − CF3CO2H). Calcd for C10H12F4N2O4S
(MW = 332.27): C, 36.15; H, 3.64; N, 8.43; F, 22.87; S, 9.65.
Found: C, 35.91; H, 3.77; N, 8.05; F, 22.99; S, 9.19.
3-((3R,4R)-3-Fluoro-1-methylpiperidin-4-yl)-
thiazolidine-2,4-dione (24). (3R,4R)-23·TFA salt (518.0 g,
H
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
1.56 mol) was added to D.I. H2O (680.00 mL) at 20 °C and
stirred for 10 min; DCM (6.8 L) was added, and the triphasic
mixture was stirred for an additional 10 min. Sodium carbonate
(203.1 g, 1.92 mol) was added, and a biphasic, gummy slurry
resulted. The biphasic mixture was treated with the solution of
formaldehyde (158.3 mL; 2.11 mol) and stirred for 10 min,
followed by the addition of sodium triacetoxyborohydride
(469.2 g, 2.10 mol) portionwise over 1 h (23.46 g/portion ×
20). The resulting mixture was stirred at 20 °C for 21 h, and
the progress of the reaction was monitored by HPLC and LC/
MS. The reaction was carefully quenched with saturated
aqueous NaHCO3 (6.0 L) over 20 min (Caution: there is
significant off-gassing during the addition of the first 1.2 L). The
layers were separated once gas evolution had ceased; the
organic phase was washed with brine (3.0 L), and the combined
aqueous phase was extracted with DCM (2.0 L). The combined
organic phase was dried over Na2SO4, concentrated at 36 °C
and then 66 °C at 12 mmHg to afford 339.8 g (88% isolated
yield; 98.4% of 24; chiral HPLC area% with 99.2% de) of
(3R,4R)-24 as an off-white to slightly beige solid. Mp = 85−86
1.71, 8.56 Hz, 1 H), 2.12 (ddd, J = 4.65, 4.89, 10.0 Hz, 1 H),
2.38 (s, 3 H, CH3), 2.56 (ddd, J = 4.16, 4.40, 12.7 Hz, 1 H),
2.90 (d, J = 11.5, Hz, 1 H), 3.25−3.30 (m, 1 H), 4.36 (ddd, J =
4.16, 4.89, 11.7 Hz, 1 H), 5.37 (dddd, J = 5.13, 5.14, 10.0, 50.6
Hz, 1 H), 5.79 (s, 2 H), 6.67 (d, J = 8.56 Hz, 1 H), 7.34 (d, J =
8.80 Hz, 2 H), 7.50 (d, J = 8.80 Hz, 1 H), 7.72 (s, 1 H), 7.96 (s,
1 H), 8.01 (s, 1 H), 8.22 (s, 1 H). 13C NMR (400 MHz,
CDCl3) δ 167.67, 166.27, 139.93, 135.43, 134.25, 134.10,
133.63, 132.47, 129.72, 128.91, 128.71, 126.65, 126.40, 124.94,
124.72, 122.14, 119.42, 109.93, 85.23, 59.20, 56.88, 54.33,
48.76, 45.66, 26.42. LC/MS m/z 553.1 (MH)+. Calcd for
C25H21ClF4N4O2S + 0.023 C6H14 + 0.189 C6H5CH5 (MW =
572.52): C, 55.53; H, 4.02; N, 9.79; Cl, 6.19; F, 13.27; S, 5.60.
Found: C, 55.42; H, 3.87; N, 9.82; Cl, 6.32; F, 13.35; S, 5.56.
(Z)-5-((1-(4-Chloro-2-(trifluoromethyl)benzyl)-1H-in-
dazol-5-yl)methylene)-3-((3R,4R)-3-fluoro-1-methylpi-
peridin-4-yl)thiazolidine-2,4-dione Hydrochloride Salt
(26 HCl Salt). To a solution of compound 26 free base
(250.0 g, 0.452 mol) in THF (2.21 L), stirred at 20 °C, was
added at 20 °C hydrogen chloride (904 mL, 0.904 mol; 1 M in
Et2O) over 45 min (the internal temperature was 24.6 °C after
1.0 equiv of HCl was added), and the resulting clear solution
was stirred for 30 min. Diethyl ether (10 L) was added as a
small stream over 90 min with vigorous stirring, and the
resulting suspension was stirred at 20 °C for 30 min. The solid
was collected by filtration, washed with Et2O (60 mL × 2), air-
dried under N2 flow, and then placed in a vacuum oven at 69
°C at 10 mmHg for 22 h to afford 263.0 g (98.7% isolated yield,
99.4% of 26 HCl salt with 99.8% de; 0.29% of deschloro 27,
chiral HPLC area%) of pure 26 HCl salt as a slightly yellow,
1
°C. H NMR (300 MHz, CDCl3) δ 1.69 (br d, J = 13.3, Hz, 1
H), 2.03 (dd, J = 2.20, 7.10 Hz, 1 H), 2.07 (ddd, J = 5.38, 6.60,
10.0 Hz, 1 H), 2.34 (s, 3 H, CH3), 2.48 (ddd, J = 4.12, 4.16,
12.5 Hz, 1 H), 2.86 (d, J = 10.0, Hz, 1 H), 3.22−3.30 (m, 1 H),
3.96 (s, 2 H), 4.19 (ddd, J = 4.65, 4.89, 11.0 Hz, 1 H), 5.36
(dddd, J = 5.13, 5.14, 10.0, 50.6 Hz, 1 H). 13C NMR (400
MHz, CDCl3) δ 171.63, 171.45, 84.91, 59.11, 56.94, 54.25
45.61, 33.24, 26.11. LC/MS m/z 233.1 (MH)+. Calcd for
C9H13FN2O2S (MW = 232.28): C, 46.54; H, 5.64; N, 12.06; F,
8.18; S, 13.80. Found: C, 46.41; H, 5.54; N, 11.73; F, 8.05; S,
13.54.
1
powdery solid. Mp = 256.6 °C (DSC). H NMR (300 MHz,
(Z)-5-((1-(4-Chloro-2-(trifluoromethyl)benzyl)-1H-in-
dazol-5-yl)methylene)-3-((3R,4R)-3-fluoro-1-methylpi-
peridin-4-yl)thiazolidine-2,4-dione (26). To a solution of
compound 24 (332.1 g, 1.42 mol) in toluene (4.94 L) at 20 °C
was added aldehyde 5 (484.2 g, 1.42 mol) and was stirred for 5
min. Piperidine (42 mL, 0.4245 mol) was added, and the
reaction flask was wrapped with aluminum foil and heated to
reflux (111−113 °C) for 86 h; the progress of the reaction was
monitored by HPLC and LC/MS (until HPLC determined the
reaction mixture to be 88.2% of 26, 10.5% of starting 5 and
0.55% of deschloro 27). The reaction was cooled to 20 °C over
6 h (and a yellowish solid precipitated below 30 °C). The slurry
was cooled further to 0 °C over 2 h and stirred for an additional
hour. The solid was collected by filtration, washed with toluene
(200 mL) and then hexanes (200 mL × 2), dried by air-suction
for 1 h, and then placed in a vacuum oven at 50 °C and
protected from the light for 68 h under nitrogen flow to afford
518.0 g (66% isolated yield, 99.6% of 26 plus 0.31% of
deschloro 27; chiral HPLC area%) of 26 free base with 99.7%
de as a slightly yellow solid.
DMSO-d6) δ 2.11−2.22 (m, 1 H), 2.68 (dd, J = 10.8, 12.9 Hz, 1
H), 2.81 (s, 3 H, CH3), 3.18−3.32 (m, 1 H), 3.32−3.45 (m, 1
H),), 3.45−3.58 (m, 1 H), 4.70 (dd, J = 4.65 9.1, Hz, 1 H), 5.57
(dddd, J = 4.89, 5.13, 10.3, 49.2 Hz, 1 H), 5.87 (s, 2 H), 6.68
(d, J = 8.56 Hz, 1 H), 7.66 (dd, J = 2.20, 8.56 Hz, 2 H), 7.69
(dd, J = 1.47, 9.10 Hz, 1 H), 7.82 (d, J = 8.80 Hz, 1 H), 7.88 (d,
J = 2.20 Hz, 1 H), 8.14 (s, 1 H), 8.19 (s, 1 H), 8.38 (s, 1 H),
11.43−11.69 (br s, 1 H). 13C NMR (400 MHz, DMSO-d6) δ
167.36, 165.45, 140.06, 135.74, 134.69, 134.35, 132.88, 131.02,
128.51, 127.92, 126.14, 126.03, 124.78, 124.68, 124.21, 121.95,
118.05, 110.74, 82.57, 53.64, 53.47, 51.71, 48.20, 42.53, 22.65.
LC/MS m/z 553.1 (MH+ of 26 free base). Calcd for
C25H21Cl1F4N4O2S + 1.0 HCl + 0.16 H2O (MW = 592.51):
C, 50.44; H, 3.84; N, 9.41; Cl, 11.91; F, 12.76; S, 5.38. Found:
C, 50.34; H, 3.81; N, 9.38; Cl, 12.22; F, 12.36; S, 5.05.
AUTHOR INFORMATION
■
Corresponding Author
*Telephone: (215)628-7829. Fax: (215)540-4611. E-mail:
xli6@its.jnj.com.
The filtrate (5.6 L) was concentrated at 69 °C to 1.5 L and
then gradually cooled to 20 °C over 1 h after seeding with pure
crystalline 26 at 40 and 30 °C; the resulting yellowish
suspension was cooled further to 0 °C over 1 h and was
stirred for an additional 30 min. The solid was collected by
filtration, washed with hexanes (200 mL × 2), dried by air-
suction for 1 h, and then placed in a vacuum oven at 60 °C,
protected from the light under nitrogen flow for 20 h, to afford
an additional 126.0 g (16% isolated yield, 99.1% of 26, and
0.80% of deschloro 27; chiral HPLC area%) of 26 free base
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the following Janssen Pharmaceutical R&D,
L.L.C. co-workers: (1) Michael Kolpak and Greg Leo, of
Discovery Sciences, for helping in structural elucidation of the
compounds 28−30, (2) Steve Stefanick, of API Small
Molecules, for conducting RC1 analysis, (3) Hilde Vanbaelen,
Jef Proost, Luc Vrins, and Marc Verbeek, The Center of
1
with 99.8% de as slightly yellow solid. Mp = 167−168 °C. H
NMR (300 MHz, CDCl3) δ 1.71−1.81 (m, 1 H), 2.08 (dd, J =
I
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Preparative Separation Techniques at Beerse, Belgium, for
chiral separations of compounds 14/15 and 21/22.
REFERENCES
■
(1) (a) Bignan, G.; Cheung, W.; Gaul, M.; Huang, H.; Li, X.; Patch,
R.; Patel, S.; Player, M.; Xu, G.; Zhao, B.-P. WO/2011/149841, 2011.
(b) Li, X.; Russell, R. K.; Branum, S. T.; Spink, J. M.; Ballentine, S. A.;
Teleha, C. A.; Wells, K. M.; Beauchamp, D. A.; Patch, R. J.; Huang, H.;
Player, M.; Murray, W. 246th American Chemical Society National
Meeting, Indianapolis, IN, United States, September 8−12, 2013,
Abstract number ORGN-216.
(2) Slade, D. J.; Pelz, N. F.; Bodnar, W.; Lampe, J. W.; Watson, P. S.
J. Org. Chem. 2009, 74, 6331.
(3) Collot, V.; Dallemagne, P.; Bovy, P. R.; Rault, S. Tetrahedron.
1999, 55, 6917.
(4) (a) Patta, S. R.; Suresh, C.; Pujar, V. D.; Reddy, V. V. K.; Rasal, V.
P.; Koti, B. C. Indian J. Chem. 2005, 2404. (b) Maccari, R.; Ottana, R.;
Curinga, C.; Vigorita, M. G.; Rakowitz, D.; Steindl, T.; Langer, T.
Bioorg. Med. Chem. 2005, 13, 2809. (c) Corona, J. A.; Davis, R. D.;
Kedia, S. B.; Mitchell, M. B. Org. Process. Res. Dev. 2010, 14, 712.
(e) Chen, S.; Chen, L.; Le, N. T.; Zhao, C.; Sidduri, A.; Lou, J. P.;
Michoud, C.; Portland, L.; Jackson, N.; Liu, J.-J.; Konzelmann, F.; Chi,
F.; Tovar, C.; Xiang, Q.; Chen, Y.; Wen, Y.; Vassilev, L. T. Bioorg. Med.
Chem. Lett. 2007, 17, 2134.
́
(5) Boto, A.; Hernandez, R.; Leon, Y.; Murguía, J. R.; Rodriguez-
Afonso, A. Eur. J. Org. Chem. 2005, 673. (b) Saavedra, J. E. J. Org.
Chem. 1979, 44, 4516. (c) Bosmans, J.-P. R. M. A.; Gijsen, H. J. M.;
Mevellec, L. A. WO/2005/000838, 2005. (d) Kratzel, M.; Hiessbock,
̈
R. Heterocycles 1995, 41, 897. (e) Daly, A. M.; Dalton, C. T.; Renehan,
M. F.; Gilheany, D. G. Tetrahedron Lett. 1999, 40, 3617.
(6) (a) Cresswell, A. J.; Davies, S. G.; Lee, J. A.; Roberts, P. M.;
Rusell, A. J.; Thomson, J. E.; Tyte, M. J. Org. Lett. 2010, 12, 2936.
(b) Ready, J. M.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2002, 41, 1394.
(c) Tandon, V. K.; Chandra, A. Tetrahedron Lett. 1993, 34, 4403.
(d) Zhao, S.; Ghosh, A.; D’Andrea, S. V.; Freeman, J. P.; Von
Voigtlander, P. F.; Carter, D. B.; Smith, M. W.; Bilnn, J. R.;
Szmuszhovicz, J. Heterocycles 1994, 39, 163. (e) Imanishi, T.; Imanishi,
I.; Momose, T. Synth. Comm. 1978, 8, 99.
(7) (a) White, J. M.; Tunoori, A. R.; Turunen, B. J.; Georg, G. I. J.
Org. Chem. 2004, 69, 2573. (b) Lal, G. S.; Pez, G. P.; Pesaresi, R. J.;
Prozonic, F. M. Chem. Commun. 1999, 215. (c) Singh, R. P.; Shreeve,
J. M. Synthesis 2002, 17, 2561. Singh, R. P.; Shreeve, J. M. J. Fluorine
Chem. 2002, 116, 23.
(8) Once pure 14 and 15 were obtained, it was found that (3R,4R)-
14 is photosensitive and with a Z-/E-isomeric mixture (∼97−95%/3−
5%) after exposure to light for 20 h. It is necessary to protect (3R,4R)-
14 (and its derivatives) from the light during the reaction or when
stored.
(9) Shaw, S. J.; Goff, D. A.; Boralsky, L. A.; Irving, M.; Singh, R. J.
Org. Chem. 2013, 78, 8892.
(10) Grunewald, G. L.; Caldwell, T. M.; Li, Q.; Criscione, K. R. J.
Med. Chem. 2001, 44, 2849.
J
dx.doi.org/10.1021/op400325r | Org. Process Res. Dev. XXXX, XXX, XXX−XXX