Asymmetric synthesis of a TRPV1 antagonist via tert -butanesulfinamide- directed reductive amination with a chromanone

Article abstract of DOI:10.1021/op400184f

An expedient asymmetric synthesis of TRPV1 antagonist 1 has been developed and demonstrated on multikilogram scale. The enabling route to 1 is detailed herein and characterized by the following key transformations: an aldol-cyclodehydration sequence to install the chromanone, and an auxiliary-mediated diastereoselective reductive amination.

Full text of DOI:10.1021/op400184f

Article
pubs.acs.org/OPRD
Asymmetric Synthesis of a TRPV1 Antagonist via tert-
Butanesulfinamide-Directed Reductive Amination with a
Chromanone
Mary E. Bellizzi, Ashok V. Bhatia, Steven C. Cullen, Jorge Gandarilla, Albert W. Kruger,*
and Dennie S. Welch*
Process Research and Development, AbbVie Inc., 1 North Waukegan Road, North Chicago, Illinois 60064-1802, United States
S
* Supporting Information
ABSTRACT: An expedient asymmetric synthesis of TRPV1 antagonist 1 has been developed and demonstrated on
multikilogram scale. The enabling route to 1 is detailed herein and characterized by the following key transformations: an aldol-
cyclodehydration sequence to install the chromanone, and an auxiliary-mediated diastereoselective reductive amination.
Scheme 1. First Generation Synthesis
INTRODUCTION
■
Transient receptor potential vanilloid-1 (TRPV1) has been
identified as an important biological receptor for the reduction
of nociceptive pain and has stimulated the development of
innovative medicines from pharmaceutical companies over the
past decade.1,2 In order to facilitate further examination of 13 as
a lead candidate (Figure 1), an expedient and efficient synthesis
was required.
Figure 1. TRPV1 antagonist.
RESULTS AND DISCUSSION
■
The first generation route commenced with commercially
available 4′-chloro-2′-hydroxy acetophenone (2), which was
condensed with 1,3-difluoroacetone (DFA) to generate
chromanone 3.4 Corey-Itsuno reduction5a,b of 3 afforded the
alcohol 4 in excellent yield with high enantioselectivity (95%,
96% ee). Conversion of 4 to azide 5 with DPPA was followed
by reduction to the amine using Ra−Ni. Formation of the D-
tartaric acid salt 6 further enriched the optical and chemical
purity to >99%.6 The synthesis of 1 culminated in a phenyl
chloroformate-mediated coupling between 6 and 3-methyl-
isoquinolin-5-amine (7).
Scheme 2. Fate of Mono-Fluoro Impurity in API
level in isolated 1. In addition, azide intermediate 5 was found
to be highly energetic, making scale-up less than ideal from a
safety perspective.7
The pyrrolidine promoted aldol-cyclodehydration of 2 to 3
was re-examined to determine the origin of monofluoro
impurity 8. After ruling out the presence of 1-fluoroacetone
While the first generation route provided a reliable means to
access 1 on small scale, there were several underlying issues that
limited the scalability of this approach. Of primary concern was
the fact that the condensation of 4′-chloro-2′-hydroxy
acetophenone with 1,3-difluoroacetone was accompanied by
formation of monofluoro impurity 8 (Scheme 2). Despite
crystallization of 6 and 1, monofluoro impurities were not
rejected by the process, leading to contaminant 9 at the 3−4%
Received: July 8, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op400184f | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
in the 1,3-difluoroacetone, it was concluded that 8 was
generated under the reaction conditions. Unfortunately, an
extensive solvent, temperature, stoichiometry, and additive
screen of the one-pot condensation/cyclization did not
establish conditions to effect this transformation without
formation of 8 in appreciable quantities.
An alternative, stepwise approach to 3 via 10 was proposed
with the aim of reducing the amount of monofluoro impurity 8
formed in the sequence. To that end, a temperature screen was
conducted to identify conditions which could efficiently provide
β-hydroxy ketone 10 (Table 1).8 The application of non-
Table 2. Sulfinylimine Formation Optimization
Table 1. Aldol Reaction Survey
conc
(M)
temp
Ti(OEt)4
(equiv)
6 h
24 h
a
a
entry
solvent
THF
THF/
(°C)
(% 3)
(% 3)
1
2
0.50
0.50
70
70
2.0
2.0
23.5
86.4
10.4
69.0
EtOH
3
4
5
6
7
toluene
0.50
0.50
1.0
100
80
2.0
2.0
2.0
3.0
4.0
5.2
3.7
3.3
4.6
3.5
2.3
2-MeTHF
2-MeTHF
2-MeTHF
2-MeTHF
10.6
12.5
temp
yield
(%)
ratio 2:10 (by
total unknown impurities
a
b
entry
(°C)
HPLC)
(%)
80
b
0.50
0.50
80
ND
1
2
3
4
−15
−45
−50
−60
53
80
84
28:72
4:96
3:97
5:95
39.0
12.4
11.3
5.5
80
7.8
a
b
Refers to the peak area percent of 3 by HPLC. Not determined.
c
ND
a
The temperature at which DFA was charged to the enolate of 2.
Refers to sum total peak area percent by HPLC of compounds not
corresponding to 2 or 10. Not determined.
formation, the reaction was conducted in toluene at elevated
temperature: an improvement in reaction rate was indeed
observed at 100 °C; however, significant degradation of the
product was observed by HPLC analysis (entry 3). Switching
the solvent to 2-MeTHF at 80 °C showed a similar rate
enhancement, and the level of degradation was low compared
to that observed in toluene: only 3.3% of 3 remained after 24 h
(entry 4). The reaction rate was unaffected by a 2-fold increase
in reaction concentration (entry 5 vs entry 4). An increase in
equivalents of Ti(OEt)4 produced only a modest improvement
in reaction rate and conversion (entries 4, 6, and 7).
b
c
cryogenic conditions was accompanied by lower yields,
recovery of 2, and impurity formation (entry 1). A notable
temperature effect was apparent with this reaction; higher yields
and a decrease in total impurities were observed at lower
temperatures (entries 2−4).
Dehydration of 10 using TFAA and pyridine to give 11 could
be achieved in quantitative yield (eq 1). The enone was then
Direct reduction of the imine, without isolation, yielded
promising results.10 Exposure of the crude imine to NaBH4 at
−50 °C and warming of the solution to −10 °C afforded 15 in
88.6% de (Table 3, entry 1). A negligible decrease in
diastereoselectivity was observed in the reaction that was
carried out entirely at −10 °C (entry 2), suggesting a selective
reaction to be feasible at noncryogenic temperatures. Increasing
the Ti(OEt)4 charge from 2 to 3 equiv caused a measurable
increase in the diastereoselectivity of the reduction (entry 3). A
decrease in reaction concentration lowered the selectivity of the
reduction (entry 4 vs entry 3). While literature precedent
cited12a the presence of 2% water in the reduction of isolated
ketimine to be beneficial to the stereoselectivity, the
introduction of water into the one-pot reaction produced the
opposite effect (entry 5). The amount of ethanol present in the
imine reduction was critical for the best results. Azeotropic
removal of the EtOH led to lower diastereoselectivity and a
relatively slow reduction (entry 6). Moreover, addition of
EtOH (1 equiv) to the reaction solution also lowered selectivity
(entry 7). As such, the single equivalent of EtOH generated in
the formation of the sulfinyl imine appeared to be optimal for
the one-pot stereoselective reduction of 14.
elaborated to 3 in excellent yield by treatment with DBU in
ethanol. Further investigation revealed that isolation of the
somewhat unstable enone 11 was not necessary. Simple
addition of EtOH and DBU to the reaction solution containing
11 led to chromanone 3 upon heating to 50 °C in 92% yield,
for the two steps.9 Importantly, formation of monofluoro
impurity 8 through this sequence was limited to <0.3%.
Reductive Amination Route. As noted earlier, removal of
azide 5 from the synthetic route was also of primary interest.7
An alternative approach for asymmetric installation of the
benzylic amine, via Ellman auxiliary-directed diastereoselective
imine reduction,10,11 appeared promising on the basis of
literature precedent with structurally similar substrates (eq 2).12
Subjection of 3 to the standard conditions11 for imine
formation revealed that after 24 h more than 10% of 3
remained (Table 2, entry 1). A report on an efficient tetralone
to sulfinyl ketimine transformation led us to try a 1:1 THF/
EtOH cosolvent system, but the rate of conversion was slower
(entry 2).12b In an attempt to drive the sulfinyl ketimine
With the ideal conditions for sulfinyl imine formation and
reduction identified, the one-pot procedure was implemented
on scale (5.0 kg) to provide 15 in 80% yield and 92% de (eq 3).
Exposure of 15 to acidic methanolic conditions, in MTBE,
B
dx.doi.org/10.1021/op400184f | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
conditions, did not improve the enantiopurity. Inspection of
the X-ray powder pattern (PXRD) of the isolated crystals
showed that a new, more stable,13 nonsolvated crystal form had
been formed. The nonsolvated form showed diminished
rejection of the unwanted isomer relative to that observed
with the CH3CN solvate. Surprisingly, chromanamine 13
obtained from the azide route repeatedly yielded the D-tartrate
salt as the CH3CN solvate even when seeded with the
nonsolvated crystal form, and excellent enantiomer rejection
was maintained.
Table 3. Diastereoselectivity of the Reduction of 14 as a
Function of Reaction Conditions
conc.
(M)
Ti(OEt)4
a
de
(%)
c
entry
temp. (°C)
(equiv)
additive
These unusual observations were difficult to rationalize, but it
appears that the thermodynamics of the solid form stability was
a function of the impurity profile. The significantly less pure
matrix of material from the azide sequence seems to favor
formation of the acetonitrile solvate over the nonsolvated form.
The benefits of the sulfinyl imine route merited an
exploration of an improved salt isolation, without the inherent
complexities of the D-tartrate salt. A series of chiral acids was
examined and it was found that the L-pyroglutamic acid salt
offered the required rejection of the undesired enantiomer of
13 (eq 4, 99.6% ee). Only a single polymorph of 17 was
observed during the development and scale-up of this
crystallization.
1
2
3
4
5
6
7
0.25
0.25
0.25
0.13
0.25
0.25
0.25
−50 to −10
−10
−10
−10
−10
2.0
2.0
3.0
3.0
3.0
3.0
3.0
n/a
n/a
n/a
n/a
88.6
87.4
91.4
89.5
74.5
65.9
85.0
b
H2O (2%)
−10
−10
−EtOH
+EtOH
(1 equiv)
a
The amount of titanium(IV) ethoxide employed in the formation of
sulfinyl imine 14. Indicates the amount of water charged on a weight
basis, relative to the quantity of 2-MeTHF present. Measured by
HPLC.
b
c
resulted in cleavage of the chiral auxiliary and concomitant
precipitation of HCl salt 16 from the solution, in 90% yield.
The quality and downstream performance of chromanamine
13 depended on its mode of preparation (Scheme 3). The
The synthesis of the coupling partner for 17 is illustrated in
Scheme 4. Methylisoquinoline 20 was commercially available,
Scheme 4. Synthetic Route to 7
Scheme 3. Purity of Freebase 13, Comparison by Approach
however not in sufficient quantities to support a multikilogram
campaign. A lab-scale procedure was developed that featured a
reductive amination of 19 with benzylamine that proceeded in
93% yield.14 Exposure of the reductive amination product to
neat chlorosulfonic acid, according to the method of Kido,15,16
gave 20 in 50% yield. This 2-step preparation of 20 was
amenable to scale-up to 5 kg at an external vendor.
relatively higher purity of 13 that was generated from sulfinyl
imine can be attributed to the inherently cleaner reaction
profile and the crystalline intermediate (16). The azide
chemistry generally provided crude 13 in approximately 65−
75 area% and 95% ee; isolation of the 1:1 D-tartrate salt of 13
(CH3CN solvate) improved the purity to 99.5 area% and 99.2%
ee.
Nitration17 of 20 gave 21 in 68% yield, along with 4.8% of a
regioisomer after isolation (91.2:8.8 ratio in the crude reaction
mixture).18 Purification of 21 by crystallization afforded variable
results at this stage. After reduction of both 21 and 22 with 5%
Pd/C in EtOH/THF, the desired amine 7 was isolated by
crystallization and contained less than 0.05% of the
regioisomer.
The sulfinyl imine route afforded material with 96 area%
purity, much higher than the azide route, albeit with lower
enantiomeric purity (92% ee). Interestingly, preparation of D-
tartrate salt of 13 from the sulfinyl imine route, under identical
The coupling to produce 1 was achieved in 80% yield by
activation of the 3-methylisoquinolin-5-amine (7) with phenyl
chloroformate followed by addition to a mixture of
pyroglutamate salt 17 and Hunig’s base (eq 5).
̈
C
dx.doi.org/10.1021/op400184f | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
1,3-Difluoroacetone (2.78 kg, 29.6 mol, 1.2 equiv) was charged
over 35 min at <−50 °C. After 10 min, HPLC analysis19
indicated 98% conversion of 4-chloro-2-hydroxyacetophenone
to 10. The reaction mixture was warmed to −20 °C and
quenched into a 10 w/w% aqueous KH2PO4 solution (181 kg).
Ethyl acetate (77 kg) was added and the layers were separated.
The organic layer was washed sequentially with 10 w/w%
aqueous KH2PO4 (91 kg), then 20 w/w% aqueous NaCl (60
kg). The organic layer was concentrated under reduced
pressure to ∼55 L. Ethyl acetate (50 kg) was charged and
the mixture was distilled to ∼35 L. Pyridine (25 kg) was then
added and the distillation was continued until the volume was
∼23 L. HPLC analysis of the pyridine solution indicated that
the solution was 23.9 w/w% 10 and 88.5 area% (Assay = 5.46
kg 10, 84% yield). Characterization of the isolated product: 1H
NMR (400 MHz, CDCl3) δ 12.00 (s, 1H), 7.73 (d, J = 8.7 Hz,
1H), 7.06 (d, J = 2.0 Hz, 1H), 6.95 (dd, J = 8.7, 2.1 Hz, 1H),
4.58 (d, J = 2.1 Hz, 2H), 4.46 (d, J = 2.1 Hz, 2H), 3.90 (s, 1H),
3.34 (t, J = 1.6 Hz, 2H); 13C NMR (101 MHz, CDCl3) δ
203.94, 162.88, 143.07, 131.02, 119.85, 118.61, 117.92, 84.86
(d, J = 3.9 Hz), 83.11 (d, J = 3.9 Hz), 72.69 (t, J = 18.1 Hz),
38.42 (t, J = 3.6 Hz); IR (KBr) 3484, 2965, 1635, 1565, 1019
cm−1. HPLC ret time: 8.41 min. Anal. Calcd for C11H11ClF2O3:
C, 49.92; H, 4.19; N, 0.0. Found: C, 49.87; H, 4.14; N, 0.1.
7-Chloro-2,2-bis(fluoromethyl)-chroman-4-one (3). A
vessel was charged with a solution of 10 (22.8 kg of a 23.9 w/w
% solution in pyridine, 20.4 mol) and pyridine (24.3 kg). The
solution was cooled to 0 °C and trifluoroacetic anhydride (5.7
kg, 27.8 mol, 1.35 equiv) was added over 1.25 h at <5 °C. After
45 min, HPLC analysis19 indicated 99.4% conversion of 10 to
11. Ethanol (33 kg) and DBU (9.2 kg, 61.8 mol, 3 equiv) were
added sequentially over 15 min at <5 °C. The reaction mixture
was warmed to 50 °C. After 8 h, HPLC analysis indicated
97.8% conversion of 11 to 3. The reaction mixture was
concentrated under reduced pressure to ∼60 L to remove the
bulk of the ethanol, then MTBE (60 kg) and a solution of 2 N
aqueous HCl (66 kg) were added. The organic layer was
separated and washed with 2 M aqueous NaOH (60 kg)
followed by a 10 w/w% solution of aqueous KH2PO4 (60 kg).
The organic layer was concentrated under reduced pressure to
∼26 L, then 2-MeTHF (54 kg) was added and distillation
followed to achieve a volume of ∼40 L. 2-MeTHF (53 kg) was
added and the product-containing organic layer was concen-
trated under reduced pressure to ∼26 L. HPLC analysis
indicated 4.37 kg of 3 in the 2-MeTHF solution (87% yield).
The final product 1 could be isolated directly from the
reaction mixture by the addition of water. The levels of 7
remaining after the first isolation were variable. Thus, a
recrystallization from acetone/water was developed to reject
remaining 7. Indeed, even when 7 was spiked to the crude
mixture at 16 mol%, the level of 7 in isolated 1 was reduced to
3 ppm. The crystallization also delivered reproducible form and
particle size.
CONCLUSIONS AND SUMMARY
In summary, a multikilogram enabling route to a novel TRPV1
antagonist has been developed (Scheme 5). The second
■
Scheme 5. Pilot-Plant Campaign to 1
generation route solved two key issues present in the initial
synthesis. The level of the monofluoro impurity 9 was
decreased from 4% to 0.2% and the energetic azide
intermediate was eliminated from the chemical sequence. The
chemistry has been demonstrated on multikilogram scale with a
pilot plant campaign yielding 2.6 kg 1 (>99.9% ee).
1
Characterization of the isolated product: H NMR (400 MHz,
DMSO) δ 7.72 (d, J = 8.4 Hz, 1H), 7.24 (d, J = 2.0 Hz, 1H),
7.13 (dd, J = 8.4, 2.0 Hz, 1H), 4.75 (d, J = 3.5 Hz, 2H), 4.64 (d,
J = 3.5 Hz, 2H), 3.02 (s, 1H); 13C NMR (101 MHz, CDCl3) δ
187.49, 158.62, 142.24, 127.56, 122.50, 118.26, 118.06, 82.80
(d, J = 4.7 Hz), 81.02 (d, J = 4.7 Hz), 80.64 (t, J = 18.4 Hz),
38.15 (t, J = 3.2 Hz); IR (KBr) 3081, 1702, 1602, 1570, 1422,
1033 cm−1. HPLC ret time: 10.51 min. Anal. Calcd for
C11H9ClF2O2: C, 53.57; H, 3.68; N, 0.0. Found: C, 53.98; H,
3.36; N, <0.05; mp 59−61 °C.
(R)-N-((R)-7-Chloro-2,2-bis(fluoro-methyl)chroman-4-
yl)-2-methylpropane-2-sulfinamide (15). A vessel was
charged with a solution of 3 (29.5 kg, 14.8 w/w% solution in
2-MeTHF), (R)-(+)-2-methyl-2-propanesulfinamide (2.4 kg,
19.8 mol, 1.1 equiv), 2-MeTHF (6 kg) and Ti(OEt)4 (12.4 kg,
54.4 mol, 3.0 equiv). The reaction solution was warmed to 80
°C. After 20 h, HPLC analysis20 indicated 5.4% remaining 3.
This solution was charged to a suspension of NaBH4 (1.4 kg,
EXPERIMENTAL SECTION
■
HPLC data were collected using an Agilent 1100 or 1200 series
HPLC. Chromatographic conditions are reported as part of the
experimental descriptions below. Retention times are un-
corrected. Assay yields were obtained by HPLC using pure
compounds as standards. Isolated yields refer to yields
corrected for purity on the basis of HPLC assays using purified
standards.
1-(4-Chloro-2-hydroxyphenyl)-4-fluoro-3-(fluoro-
methyl)-3-hydroxybutan-1-one (10). A vessel was charged
with THF (48.7 kg), diisopropylamine (5.2 kg, 51.7 mol, 2.1
equiv), and then cooled to 0 °C. Butyl lithium (14.3 kg, 51.7
mol, 2.1 equiv) was added over 40 min at <10 °C. After 15 min,
a solution of 4′-chloro-2′-hydroxy acetophenone (2) (4.2 kg,
24.6 mol) in THF (3.6 kg) was added over 25 min at <10 °C.
After 15 min at 0 °C, the solution was then cooled to −50 °C.
D
dx.doi.org/10.1021/op400184f | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
37 mol, 2.0 equiv) and THF (30 kg) over 60 min at <−10 °C.
After 18 h, HPLC analysis indicated consumption of the
sulfinyl ketimine intermediate. The reaction mixture was then
warmed to 2 °C and diluted with 2-MeTHF (64 kg). The
reaction was quenched with the cautious addition of a 14.1 w/w
% aqueous solution of sodium glycolate/glycolic acid (145 kg,
15 equiv relative to substrate and prepared by the addition of
NaOH (0.8 equiv relative to the glycolic acid) to a 15 w/w%
aqueous solution of glycolic acid; final pH measured to be 4.7)
over 40 min at <10 °C. The mixture was warmed to 20 °C and
aged with mixing for 12 h. The resulting layers were separated.
The organic layer was then mixed with an aqueous sodium
glycolate/glycolic acid solution (73 kg) for 1 h. The resulting
layers were separated. The organic layer was mixed with water
(55 kg) for 30 min and then allowed to stand for 15 min. The
resulting layers were separated. The organic layer was agitated
with a 10 w/w% aqueous solution of NaCl (55 kg) for 30 min
and allowed to settle for 30 min. The resulting layers were
separated and the organic layer distilled to ∼40 L and
codistilled with 2-MeTHF (100 kg). This solution was then
passed through a polishing filter and distillation was continued
to reach ∼60 L. A continuous distillation was performed with
MTBE (150 kg) which gave the desired product 15 (4.99 kg,
79.5% yield of the desired product diastereomer, 90.7% de, 84.0
area%) as a 7.5 w/w% solution in MTBE. Characterization of
the isolated product: NMR (600 MHz, CDCl3) δ 7.50 (dd, J =
8.3, 1.0 Hz, 1H), 6.97 (dd, J = 8.3, 2.1 Hz, 1H), 6.92 (d, J = 2.1
Hz, 1H), 4.65−4.4 (m, 5H), 3.41 (d, J = 5.8 Hz, 1H), 2.27 (dt,
J = 14.3, 5.8, 1.5 Hz, 1H), 2.14 (ddd, J = 14.3, 7.8, 2.1 Hz, 1H),
1.25 (s, 9H); 13C NMR (126 MHz, CDCl3) δ 153.2, 135.1,
130.2, 122.2, 120.7, 117.7, 82.8 (dd, J = 131.7, 3.8), 81.65 (dd, J
= 131.7, 3.8), 76.7 (t, J = 18.9), 56.1, 47.6, 30.8 (t, J = 2.7),
22.6; IR 3179, 2966, 1604, 1026 cm−1. HPLC ret time: 12.36
min. HRMS (ES+) calcd for C15H20ClF2NO2S, 351.08713,
was filtered through a polypropylene filter, and the filtrate was
transferred to a separatory funnel. The lower layer was
removed, and the organic layer was washed with a 25 w/w%
aqueous solution of NaCl (20 kg × 2). The organic layer was
then passed through an 8 in. Cuno filter housing with a Zeta
Carbon Filter 53SP and rinsed with MTBE (2 × 8 kg). The
combined filtrate was concentrated under reduced pressure to
∼5 volumes (20 L), and then codistilled with isopropyl acetate
(∼70 kg) until the water content was determined to be <0.5 w/
w% by Karl Fischer analysis. The solution was then filtered. A
separate vessel was charged with L-pyroglutamic acid (2.0 kg)
and ethanol (28 kg, 200 proof). To the freebase solution was
charged ethanol (20.0 kg) and a portion of the L-pyroglutamic
acid solution in ethanol (1.23 kg, 5% of the total solution
needed, 0.05 equiv). To the clear solution was charged 17 (20.0
g, 0.1% seed load). The remainder of the L-pyroglutamic acid
solution (26.0 kg, to make ∼1.05 equiv) was added to the thin
slurry over 4 h. The slurry was then cooled to 0 °C and the
supernatant was analyzed by HPLC to confirm appropriate
level of 13. The product was isolated by filtration and washed
with ethanol (2 × 12 kg). The crystallization provided 17 as a
beige solid (4.58 kg, 94% yield, 64.3 w/w% 13, 99.2 area%,
99.6% ee, 0.35 area% desfluoro impurity). 1H NMR (400 MHz,
DMSO) δ 7.80 (s, 1H), 7.55 (dd, J = 8.3, 0.8 Hz, 1H), 7.00
(dd, J = 8.3, 2.1 Hz, 1H), 6.89 (d, J = 2.1 Hz, 1H), 4.73−4.45
(m, 4H), 4.05−3.95 (m, 2H), 2.35−2.15 (m, 2H), 2.13−2.06
(m, 2H), 1.99−1.89 (m, 1H), 1.77 (ddd, J = 15.8, 8.0, 1.5 Hz,
1H); 13C NMR (126 MHz, DMSO) δ 176.85, 174.95, 152.91,
132.31, 129.00, 124.45, 120.75, 116.12, 83.8 (dd, J = 174.3, 5.2
Hz), 81.8 (dd, J = 174.3, 5.9 Hz), 77.53 (t, J = 17.9 Hz), 55.53,
41.84, 30.97 (t, J = 2.9 Hz), 29.30, 24.86; IR 3280, 2781, 2204,
1671, 1634, 1543 cm−1. HPLC ret time: 11.13 min. Anal. Calcd
for C16H19ClF2N2O4: C, 51.00; H, 5.08; N, 7.43. Found: C,
23
51.04; H, 5.01; N, 7.33; mp 167−169 °C; [α]D = −20.98°
(50/50 H2O/Acetonitrile, 1.17 g/100 mL).
23
Found 352.09484 [M+H]; mp 120−122 °C [α]D = −12.75°
(CHCl3, 1.073 g/100 mL).
3-Methylisoquinoline (20). A vessel was charged with
benzylamine (25.5 mL, 233 mmol, 1.0 equiv), 1,1-dimethox-
ypropan-2-one (27.6 mL, 233 mmol 1.0 equiv), and DCE (819
mL). Sodium triacetoxyborohydride (69.2 g, 327 mmol, 1.40
equiv) was then added in one portion and the reaction was
aged overnight. GC analysis21 showed complete conversion and
the reaction was diluted with a 2.5 w/w% aqueous solution of
NaHCO3 (700 mL) The resulting biphasic mixture was mixed
for 30 min. The organic layer was discarded and the aqueous
layer was basified by the addition of a 1 M NaOH aqueous
solution to a pH of ∼14. The aqueous layer was then extracted
with EtOAc (2 × 250 mL). The combined organic layers were
washed with a 10 w/w% aqueous solution of KH2PO4 to reach
a pH of ∼7−8, and then washed with a 5 w/w% aqueous
solution of NaCl. The organic layer was dried over MgSO4,
filtered, and concentrated under reduced pressure to give N-
benzyl-1,1-dimethoxypropan-2-amine as a yellow oil (93%
yield). A vessel was charged with chlorosulfonic acid (48.0 mL,
717 mmol). N-Benzyl-1,1-dimethoxypropan-2-amine (15 g,
71.7 mmol) was then added at <20 °C. Upon complete
addition the internal temperature of the reaction adjusted to
100 °C and held for 10 min. The reaction was then quenched
onto ice and MTBE was added. The organic layer was
discarded and the aqueous layer was cooled and basified by the
addition of a 50 w/w% aqueous solution of NaOH to reach pH
∼14. CH2Cl2 (200 mL) was then added and the layers were
separated. The aqueous layer was extracted with CH2Cl2 (200
mL) and the combined organic layers were washed with brine,
(R)-7-Chloro-2,2-bis(fluoromethyl)-chroman-4-ami-
nium Chloride (16). A vessel was charged with MeOH (45
kg) and then acetyl chloride (24.5 kg) was charged over 1 h at
<10 °C (Caution: Exothermic reaction). After 30 min, this
solution (28.3 mol, 1.5 equiv of HCl) was added to the
sulfinamide 15 (4.99 kg, 14.2 mol, 1.0 equiv) in MTBE (44 kg)
at 20 °C. After 7 h, HPLC analysis20 indicated complete
consumption of sulfinamide 15. The reaction mixture was
filtered and the cake was washed with MTBE (34 kg, pre-
cooled to 0 °C) to provide 16 as a tan, crystalline solid (4.14
1
kg, 93% yield, 98.5 area%, 92.2% ee). H NMR (400 MHz,
DMSO) δ 9.10 (bs, 3H), 7.76 (dd, J = 8.4, 0.4 Hz, 1H), 7.11
(dd, J = 8.4, 2.1 Hz, 1H), 7.02 (d, J = 2.2 Hz, 1H), 4.73−4.51
(m, 5H), 2.54 (dd, J = 13.7, 6.2 Hz, 1H), 2.06 (ddd, J = 14.4,
12.0, 3.3 Hz, 1H); 13C NMR (126 MHz, DMSO) δ 152.6,
133.4, 128.3, 120.8, 117.7, 116.4, 83.5 (d, J = 173.8, 5.7), 81.3
(d, J = 173.8, 5.7), 77.0 (t, J = 17.8), 41.6, 27.1; IR 2964, 2884,
2017, 1608 cm−1. HPLC ret time: 7.54 min. HRMS (ES+)
calcd for C11H12ClF2NOS, 247.05755, Found 248.06516 [M
23
+H]; mp 206−208 °C; [α]D = −9.85° (MeOH, 1.089 g/100
mL).
(R)-7-Chloro-2,2-bis(fluoromethyl)-chroman-4-amine
(S)-5-oxopyrrolidine-2-carboxylate (17). A vessel was
charged with 16 (4.1 kg salt (3.32 kg freebase), 14.4 mol,
92.2% ee, 98.5 area%), MTBE (30.6 kg), and water (41.5 kg). A
50 w/w% aqueous solution of NaOH (2.3 kg, 4 equiv) was
charged to the reactor at 20 °C. Upon dissolution, the mixture
E
dx.doi.org/10.1021/op400184f | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
dried over MgSO4, and filtered. Concentration of the organic
layer under reduced pressure provided 20 as a tan solid (5.1 g,
100 w/w%, 50% yield). The NMR spectrum of the title
compound was consistent with literature precedent.22
(R)-1-(7-Chloro-2,2-bis(fluoromethyl)-chroman-4-yl)-
3-(3-methylisoquinolin-5-yl)urea Hydrate (1). A vessel
was charged with 7 (1.2 kg, 7.59 mol, 1.0 equiv), pyridine
(0.322 kg, 4.07 mol, 0.5 equiv), DMF (20.2 kg), and
acetonitrile (10.0 kg). A solution of phenyl chloroformate
(1.422 kg, 9.08 mol, 1.2 equiv) in acetonitrile (4.7 kg) was
added to the reactor over ∼3 h at 25 °C. The reaction was
allowed to proceed overnight (16 h) at room temperature. An
in-process sample indicated 0.17 area% of 7 remained. L-
Pyroglutamate salt 17 (2.86 kg, 7.59 mol, 1.0 equiv) was then
charged to the slurry, followed by diisopropylethylamine (3.1
kg, 24.8 mol, 3.2 equiv). Analysis by HPLC24 indicated 0.1 area
% activated intermediate remaining. The reaction was diluted
with acetonitrile (10 kg). To the product-containing solution
was charged water (16 kg), followed by seeds of 1. Water (74
kg) was then charged over 6 h to the slurry. The slurry was
assayed for product in solution; a product concentration of 0.06
w/w% 1 was observed in the supernatant. The slurry was then
filtered, washed with 50 v/v% acetonitrile/water (16 kg), and
dried to provide 1 as a beige solid (2.99 kg, 89.2% yield, 97.8
w/w%, 99.4 area%). A vessel was charged with 1 and acetone
(38 kg, 15 mL/g 1) to yield a solution, which was then filtered
through a Cuno R53SP carbon pad. The pad was then washed
with acetone (7 kg). To the product-containing solution was
charged water (8 kg). To the solution was charged seeds of 1
and then water (60 kg) was added over 4 h. The resulting slurry
was sampled to check product in solution (0.04 w/w% 1 in
solution). The slurry was filtered and washed with 50 v/v%
acetone/water (2 × 15 kg) to provide 1 as an off-white solid
(2.735 kg, 90% yield, 101.3% w/w, 99.4 area%, >99.9% ee). 1H
NMR (400 MHz, DMSO) δ 9.18 (s, 1H), 8.66 (s, 1H), 8.26
(dd, J = 7.7, 1.0 Hz, 1H), 7.74 (d, J = 0.6 Hz, 1H), 7.71 (dd, J =
8.1. 1.1 Hz, 1H), 7.52 (t, J = 7.9 Hz, 1H), 7.38 (dd, J = 8.3, 0.9
Hz, 1H), 7.06 (dd, J = 8.3, 2.1 Hz, 1H), 7.03 (d, J = 8.1 Hz,
1H), 6.99 (d, J = 2.1 Hz, 1H), 5.11−4.99 (m, 1H), 4.73 (dd, J =
8.8, 1.5 Hz, 2H), 4.60 (dd, J = 8.8, 1.5 Hz, 2H), 2.66 (s, 3H),
2.37 (dd, J = 13.7, 6.0 Hz, 1H), 2.02 (ddd, J = 13.6, 10.6, 2.9
Hz, 1H); 13C NMR (101 MHz, DMSO) δ 154.6, 152.8, 151.5,
150.3, 133.1, 132.1, 128.6, 127.9, 126.4, 125.9, 122.3, 121.1,
120.6, 119.3, 116.0, 111.5, 83.5 (d, J = 151.4, 5.1), 81.8 (d, J =
151, 5.6 Hz), 77.5 (t, J = 17.7 Hz), 41.1, 29.4 (t, J = 3.4 Hz),
24.4; IR 3289, 3063, 1082, 1566, 1233 cm−1. HPLC ret time:
14.78 min. HRMS (ES+) calcd for C22H20N3O2F2Cl,
431.12854 [M+H]. IR: 1681.8, 1564.1, 1232.0, 756.7; mp
3-Methyl-5-nitroisoquinoline (21). A vessel was charged
with sulfuric acid (44 L) and the internal temperature was
adjusted to 0 °C. 20 (4.4 kg, 30.8 mol, 1.0 equiv) was charged
to the acid solution. Potassium nitrate (3.42 kg, 34.2 mol, 1.1
equiv) was added in 4 equal portions at <5 °C. The reaction
mixture was warmed to room temperature and after 90 min
HPLC analysis23 indicated the reaction to be complete (0.02
area% starting material remaining). Water (66 kg) was added to
the mixture at 0 °C. Two-thirds of the contents of the round-
bottom flask was transferred to a separatory funnel and held
while the remaining contents were processed: an aqueous 50%
w/w solution of NaOH (23.2 kg) was added to the vessel at
<60 °C, bringing the pH to ≥12. Water (30 kg) was added and
the internal temperature was adjusted to 50 °C. The mixture
was agitated for 1 h. The solids were filtered and washed with
water (20 kg). This crystallization/filtration procedure was
repeated two more times on the remaining reaction mixture to
give 21 as a tan solid (5.36 kg, 85.9 w/w%, 89 area%, 79.6%
yield). This crude product and MeOH (36.4 kg) were then
charged to a vessel. The internal temperature was then adjusted
to 45 °C. The methanol solution was then filtered through a
Cuno housing fitted with a Zeta Carbon Filter 53SP. Water (45
kg) was added to the filtrate in over 1 h and the resulting solids
were filtered and rinsed with MeOH/water (20 kg, 1:1 v/v) to
give 21 as an off-white solid (4.21 kg, 98 w/w%, 95 area%, 73%
yield). The NMR spectrum of the title compound was
consistent with literature precedent.17
3-Methylisoquinolin-5-amine (7). A vessel was charged
with 21 (4.21 kg, 22.4 mol), and 5% Pd/C Johnson Matthey
A102023−5 (178 g catalyst, lot contains 52.6% water. Catalyst/
substrate ratio: 2 w/w% on a dry catalyst basis). The reactor
was purged with nitrogen and a mixture of EtOH (50 kg) and
THF (56.5 kg) was added. The reactor was purged with
nitrogen followed by hydrogen and pressurized with hydrogen
to 30 psig. The mixture was vigorously agitated at 25 °C and 30
psig hydrogen for 4 h, after which time the reaction was
complete.24 The contents of the reactor were filtered through a
Cuno filter housing containing a 12 in. polyamide diatoma-
ceous earth filter and two back-up 2 μm nylon inline filters.
Approximately half (63 kg) of the filtrate was added to a vessel
and concentrated under reduced pressure to ∼11 L. Methanol
(20 kg) was added and concentration to ∼11 L was repeated.
Methanol (50 kg) was charged and the slurry was warmed to 40
°C to achieve complete dissolution of the solids. The
methanolic solution of 7 was filtered through a polypropylene
carbon filter and the filtrate was concentrated to 11 L. The
second half of the filtrate was processed in the same manner
and then combined with the first portion of the filtrate. The
combined filtrates were concentrated to ∼31 L. A continuous
distillation was performed using EtOAc (153 kg) and the
solution temperature was adjusted to 50 °C and held at that
temperature for 30 min to achieve full dissolution of the solids.
The solution was cooled to 0 °C over 90 min. The resulting
slurry was filtered and the wetcake was washed with chilled
ethyl acetate (18 kg) and dried to give 7 as an off-white solid
(2.34 kg, 101.2 w/w%, 100 area%, 67% yield). The NMR
spectrum of the title compound was consistent with literature
precedent.17
23
144−146 °C; [α]D = +1.17°(MeOH, 0.9 g/100 mL).
ASSOCIATED CONTENT
* Supporting Information
■
S
Spectral information for some compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: albert.kruger@abbvie.com.
*E-mail: dennie.welch@abbvie.com.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The design, study conduct, analysis, and financial support of the
study were provided by AbbVie. The authors thank Yong Chen
and Brenda Mathias for analytical support. A.V.B., S.C.C., J.G.,
F
dx.doi.org/10.1021/op400184f | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(19) HPLC method: Halo-C18 (0.46 × 150 mm, 2.7 μm); 35 °C, 1.0
mL/min; 90/10 to 10/90 0.1% aqueous HClO4/acetonitrile over 10
min, hold 6 min.
A.W.K., and D.S.W. are employees of AbbVie Inc. M.E.B. is a
former AbbVie employee.
(20) HPLC method: Zorbax Eclipse XDB-C18 (0.46 × 150 mm, 5.0
μm); 35 °C, 1.0 mL/min; 90/10 to 10/90 0.1% aqueous HClO4/
acetonitrile over 15 min, hold 5 min.
(21) GC method: Performed on an Agilent 6890 Series GC using an
HP-5 5% Phenyl Methyl Siloxane column (30.0 m × 320 μm × 0.25
μm). Method information: Injection temp = 250 °C, Column start
temp = 50 °C, method parameters: 50 to 200 °C at 10 °C/min, then
hold at 200 °C for 10 min.
(22) Balkau, F.; Heffernan, M. L. Aust. J. Chem. 1971, 24, 2311.
(23) HPLC method: Halo-C18 (0.46 × 150 mm, 2.7 μm); 45 °C, 1.5
mL/min; 90/10 to 70/30 0.1% aqueous HClO4/acetonitrile over 3
min, hold 3 min.
(24) HPLC method: Zorbax Eclipse SB (0.46 × 150 mm, 3.5 μm);
30 °C, 1.0 mL/min; 95/5 to 10/90 0.1% aqueous HClO4/acetonitrile
over 25 min, hold 5 min.
REFERENCES
■
(1) For reviews, see: (a) Kym, P. R.; Kort, M. E.; Hutchins, C. W.
Biochem. Pharmacol. 2009, 78, 211. (b) Voight, E. A.; Kort, M. E.
Expert Opin. Ther. Pat. 2010, 20, 1107. (c) Szallasi, A.; Appendino, G.
J. Med. Chem. 2004, 47, 2717.
(2) Gunthorpe, M. J.; Chizh, B. A. Drug Discov. Today 2009, 14, 56.
(3) Reilly, R. M.; McDonald, H. A.; Puttfarcken, P. S.; Joshi, S. K.;
Lewis, L.; Pai, M.; Franklin, P. H.; Segreti, J. A.; Neelands, T. R.; Han,
P.; Chen, J.; Mantyh, P. W.; Ghilardi, J. R.; Turner, T. M.; Voight, E.
A.; Daanen, J. F.; Schmidt, R. G.; Gomtsyan, A.; Kort, M. E.; Faltynek,
C. R.; Kym, P. R. J. Pharmacol. Exp. Ther. 2012, 342, 416.
(4) Kabbe, H. J. Synthesis 1978, 886.
(5) (a) Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am. Chem. Soc. 1987,
109, 5551. (b) Hirao, A.; Itsuno, S.; Nakahama, S.; Yamazaki, N. J.
Chem. Soc., Chem. Commun. 1981, 315.
(6) Voight, E. A.; Daanen, J. F.; Hannick, S. M.; Shelat, B. H.;
Kerdesky, F. A.; Plata, D. J.; Kort, M. E. Tetrahedron Lett. 2010, 51,
5904.
(7) The heat of formation of the azide was estimated to be −771 J/g,
on the basis of reaction calorimetry (RC1), while differential scanning
calorimetry (DSC) of the azide revealed a −1141 J/g exothermic
decomposition event, onset at 127 °C.
(8) Enolization and aldol reaction of hydroxyacetophenones has been
reported: Banerji, A.; Goomer, N. C. Tetrahedron Lett. 1979, 20, 3685.
(9) The presence of 12, as a mixture of alkene isomers, was detected
during the course of this reaction, but could be converted to the
desired product by warming the reaction mixture to 50 °C.
(10) Borg, G.; Cogan, D. A.; Ellman, J. A. Tetrahedron Lett. 1999, 40,
6709.
(11) Liu, G.; Cogan, D. A.; Owens, T. D.; Tang, T. P.; Ellman, J. A. J.
Org. Chem. 1999, 64, 1278.
(12) For examples, see: (a) Colyer, J. T.; Andersen, N. G.; Tedrow, J.
S.; Soukup, T. S.; Faul, M. M. J. Org. Chem. 2006, 71, 6859. (b) Han,
Z.; Koenig, S. G.; Zhao, H.; Su, X.; Singh, S. P.; Bakale, R. P. Org.
Process Res. Dev. 2007, 11, 726.
(13) Slurry conversion experiments on pure material indicated this
new anhydrous form to be more stable under the range of acetonitrile/
water compositions that were typical for the isolation (10−20%
water). The acetonitrile solvate required >99% acetonitrile/water
composition to be competitively more stable.
(14) Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.; Maryanoff, C.
A.; Shah, R. D. J. Org. Chem. 1996, 61, 3849.
(15) Kido, K.; Watanabe, Y. Chem. Pharm. Bull. 1987, 35, 4964.
(16) Reimann, E.; Hertel, R.; Krauss. Monatsh. Chem. 2008, 139,
673. Hertel describes therein a cryogenic variant of the chlorosulfonic
acid cyclization that afforded a slightly higher yield (69%).
(17) Perner, R. J.; DiDomenico, S.; Koenig, J. R.; Gomtsyan, A.;
Bayburt, E. K.; Schmidt, R. G.; Drizin, I.; Zheng, G. Z.; Turner, S. C.;
Jinkerson, T.; Brown, B. S.; Keddy, R. G.; Lukin, K.; McDonald, H. A.;
Honore, P.; Mikusa, J.; Marsh, K. C.; Wetter, J. M.; St. George, K.;
Jarvis, M. F.; Faltynek, C. R.; Lee, C.-H. J. Med. Chem. 2007, 15, 3651.
(18) Regiosomer level based on HPLC area%. By inspection of the
crude NMR, the regioisomer was tentatively identified as 22.
G
dx.doi.org/10.1021/op400184f | Org. Process Res. Dev. XXXX, XXX, XXX−XXX