Full text of DOI:10.1021/op200175t

ARTICLE
pubs.acs.org/OPRD
Asymmetric Synthesis of LFA-1 Inhibitor BIRT2584 on Metric Ton Scale
Xiao-Jun Wang,*,† Rogelio P. Frutos,*,† Li Zhang,† Xiufeng Sun,† Yibo Xu,† Thomas Wirth,‡ Thomas Nicola,‡
Lawrence J. Nummy,† Dhileep Krishnamurthy,† Carl A. Busacca,† Nathan Yee,† and Chris H. Senanayake†
†Chemical Development, Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, Connecticut 06877, United States
‡Department of Production and Engineering, Boehringer Ingelheim GmbH & Co.KG, 55216 Ingelheim am Rhein, Germany
S Supporting Information
b
ABSTRACT: The synthesis of LFA-1 inhibitor BIRT2584 on metric-ton scale was accomplished by means of a safe and robust
process. Highlights of the process include the asymmetric synthesis of the key advanced intermediate by implementation of
Seebach’s self-regeneration of stereocenters principle, and a Ph3PCl2-induced dehydration of a critical urea followed by a
regioselective bromination to give the elaborated 1H-imidazo[1,2-a]imidazol-2-one. A sulfonyl chloride intermediate was produced
through Br/Mg exchange of iodoimidazole followed by addition to SO2 in THF and subsequent oxidation. In a one-pot operation,
the sulfonyl chloride was directly reacted with L-alaninamide using NaOH as base in aqueous DMF/THF to give BIRT2584.
’ INTRODUCTION
high diastereoselectivity using Seebach’s protocol and at a fraction of
the cost of an analogous pivalaldehyde-derived template.7 The
synthesis of an isobutyraldehyde-derived template proceeded by
the synthesis of intermediate 4. Imidazolidinone 4 was initially
obtained upon condensation of 3 with isobutyraldehyde in toluene
with simultaneous azeotropic distillation of water. The above
condensation afforded a mixture of both the cis- and trans-
isomers of 4, however, the pure trans-isomer 4 was obtained by
a crystallization driven resolution in heptane in which the isomers
interconverted in solution but the trans-isomer crystallized out of
the mixture. This procedure was carried out successfully for the
synthesis of 4 in multikilogram quantities. A subsequent, more
streamlined process was developed in which the azeotropic
distillation of water was performed with EtOAc instead of toluene.
The use of EtOAc shortened the cycle time for the solvent
exchange to heptane (<5% by volume of residual EtOAc), and
the desired 4 crystallized in 75À79% yield in heptane at 50 °C.
Although 4 could be isolated in a promising 75À79% yield
directly from the reaction mixture, approximately 20% of 4 and
its cis-isomer 4a remained in the filtrate as a 1:1 mixture. As a
result, we sought to develop a practical process to recover 4 and
4a by “recycling” the mother liquor, considering it contained no
significant amount of impurities. As illustrated in Scheme 3, a
more efficient synthesis of 4 was implemented by reusing the
mother liquor from the initial reaction after collecting the first
portion of 4. Thus, the filtrate of the heptane solution from the
first batch was added to the concentrate of the second batch (the
same reaction at the same scale) after azeotropic distillation of
water in EtOAc. A solvent exchange to heptane was then
performed by distillation until the level of residual EtOAc in
heptane was reduced to <5% by volume. The desired trans-
isomer 4 was isolated in 95% yield for the second batch, and this
process was repeated for up to four cycles as summarized in
BIRT2584 is a small-molecule lymphocyte function-associated
antigen-1 (LFA-1) inhibitor under investigation as a therapeutic
agent in the treatment of immune ailments such as psoriasis,
Crohn’s disease, and rheumatoid arthritis.1 To support late-stage
development activities as well as clinical studies, the synthesis of
BIRT2584 active pharmaceutical ingredient (API) in amounts
exceeding one metric ton was needed, and a suitable scalable
process had to be developed, optimized, and implemented. The
previous syntheses of BIRT2584 followed approaches to struc-
turally related LFA-1 inhibitors as shown in Scheme 1. Earlier
routes proceeded through intermediates B or C and started from
the pivalaldehyde-derived Seebach’s template A.2 Pivalaldehyde
was a significant contributor to the overall API cost, and its
commercial sources were limited. The conversion of B to inter-
mediate D for the first route required a stoichiometric amount of
CuCl that was not cost-effective.3 The conversion of C to the
common intermediate D for the second route required an excess
of Me3Al, which posed a safety concern.4 Furthermore, the
transformation of intermediate D to iodo-imidazole E involved
iodination of a phosphate intermediate with HI, that produced
side products that needed to be removed.5 Although the previous
two approaches were employed to produce up to 10 kg of BIRT2584
for early development activities, they were not sufficient to
support a long-term development cycle. Herein, we report the
optimization and implementation of a new process for the
synthesis of LFA-1 inhibitor BIRT2584 on metric ton scale.6
’ RESULTS AND DISCUSSION
The BIRT2584 process began with the synthesis of inter-
mediate 4 from Boc-D-alanine (1) as shown in Scheme 2. Boc-D-
alanine was treated with isobutyl chloroformate/triethylamine
and 3,5-dichloroaniline was added to the resulting mixed anhy-
dride to give 2. The crude 2 was hydrolyzed with p-toluenesul-
fonic acid in toluene to afford 3 as the p-TsOH salt in 85% yield
from 1. We had previously reported that an isobutyraldehyde
derived imidazolidinone template (i.e., 5) could be alkylated with
Special Issue: Asymmetric Synthesis on Large Scale 2011
Received: June 29, 2011
Published: August 03, 2011
r
2011 American Chemical Society
1185
dx.doi.org/10.1021/op200175t Org. Process Res. Dev. 2011, 15, 1185–1191
|

Organic Process Research & Development
ARTICLE
Scheme 1. Original synthesis of BIRT2584
Table 1. Recycle of 4 in mother liquor
entry
recycle
scale (g)
yield of 4 (%)
1
2
3
4
5
initial batch
100
100
100
100
100
76
95
96
93
87
1
2
3
4
Scheme 4. Alkylation of 5 with 4-trifluoromethoxybenzyl
bromide
A highly diastereoselective alkylation of template 5 with
4-trifluoromethoxybenzyl bromide was performed next to install
the quaternary chiral center (Scheme 4). The desired diastereoi-
somer 6 was the sole product. Initially, LiHMDS was added
to a solution of 5 in THF precooled to À15 °C to preform an
enolate species. This process was exothermic and took a pro-
longed time (>2.5 h) to complete the addition of LiHMDS and
keep the reaction temperature below 0 °C on 100-kg scale at one
of our production facilities. During development activities for
this reaction on a smaller scale this addition took a shorter time,
and the stability of the enolate was not an issue. However, at
production scale, significant decomposition of the enolate was
observed over time, resulting in a moderate yield of 6 after
isolation from a mixture of IPAc/heptane.
Scheme 2. Initial synthesis of template 5
Therefore, this alkylation was further investigated as summar-
ized in Table 2. The results demonstrated that formation of
enolate at À5 to 0 °C (method A) over a longer time resulted in
more decomposition with lower recoveries of 6. The experiment
in entry 3 was a simulation of our first production run on 100-kg
scale, and the result was consistent with what we had previously
experienced. We were pleased to find that addition of LiHMDS
to a mixture of 5 and 4-trifluoromethoxybenzyl bromide in THF
at a higher temperature of À5 to 5 °C in 3 h (method B)
produced 6 in 93% yield. The rapid alkylation of the enolate as it
was formed by benzyl bromide prevented decomposition. This
procedure was implemented in our production facility and
allowed the large-scale manufacturing of 6 under noncryogenic
conditions.
Scheme 3. Recycle of 4 in mother liquor
Cleavage of the alkylated template was performed in one-pot
by the successive treatment of 6 with KOH in IPA at 60 °C
followed by 3 M sulfuric acid. Chiral amine 7 was isolated as the
p-TSA salt from acetonitrile in 87% yield. Condensation of 7 with
phenylcarbamate 88 in the presence of TEA in DMSO at 60 °C
followed by a crystallization in methanol gave urea 9 in 92% yield.
Solvents such as MeOH or CH3CN significantly slowed the
reaction. This outcome was consistent with a literature report,9
and a similar result was obtained using TEA as the solvent in the
absence of DMSO (Scheme 5). With the urea 9 in hand, we next
carried out the dehydration/cyclization to 11 which forms the
core structure of the 1H-imidazo[1,2-a]imidazol-2-one system.
Table 1. The subsequent three cycles afforded a much higher
recovery than the first. An average yield of about 90% for 4 was
achieved by this method. This procedure was carried out success-
fully and reproducibly for the synthesis of 4 on production scale.
1186
dx.doi.org/10.1021/op200175t |Org. Process Res. Dev. 2011, 15, 1185–1191

Organic Process Research & Development
ARTICLE
Table 2. Alkylation of 5 with 4-trifluoromethoxybenzyl
bromide
The subsequent bromination was performed directly on the
crude IPAc solution of 11 from the previous step which simplified
the process greatly. Similar brominations of N-alkylimidazoles,13
imidazo[1,2-a]pyrimidines,14 and imidazo[1,2-a]pyridine derivatives15
are known to be very regio- and chemoselective, but the
bromination of 11 was not straightforward and required some
optimization. Table 3 outlines the results of some typical
experiments. Treatment of 11 by slow addition of a suspension
of 1.05 equiv of NBS in IPAc gave bromoimidazole 12 in 57%
yield with only a trace amount of regioisomer 13 and <2% of
dibromide 14. A similar result was obtained when chromatogra-
phically pure 11 was used under the same conditions (entry 2,
Table 3). While the regio- and chemoselectivity were promising,
bromide 12 was obtained only in moderate yield because several
byproducts were obtained which diminished the yield by more
than 10%. Isolation and/or identification of these byproducts
were not successful since they rapidly interconverted. On the
basis of preliminary MS data, these byproducts are possibly
products of imidazole-dimerization.16 Interestingly, in the pre-
sence of Lewis acids such as PPTS and ZnBr2, the bromination
resulted in significantly increased formation of these byproducts,
accounting for up to 22% of the mixture (entries 3À4, Table 3).
These experiments suggested that the proposed dimerization of
11 may be promoted by acids, which is consistent with literature
reports of Lewis acid-promoted dimerization of imidazole-re-
lated heterocycles.17 As expected, the same bromination of 11 in
the presence of TEA significantly suppressed the formation of
dimer-derivatives (entry 5, Table 3), leading to a much improved
83% yield of 12. In this case about 1.4 equiv of NBS was needed
for a complete reaction due to the reaction between TEA and
NBS.18 Subsequently, the inorganic base K2CO3 was used to
replace TEA, and indeed, the bromination of 11 produced 12 in
86% isolated yield with no need for excess NBS. Crystallization
from IPA/water provided clean 12 and also removed the residual
Ph3PO carried over from the previous step. Controlling the
amount of NBS based on the HPLC assay of the crude IPAc
solution of 11 was critical. Excess NBS led to an increased
amount dibromide 14 and a low overall yield of BIRT2584 in the
subsequent step.
time for addition
entry
scale (kg)
methoda
of LiHMDS (h)
yield of 6 (%)b
1
2
3
4
5
0.1
0.1
A
A
A
B
B
0.5 (À10 to 0 °C)
1.5 (À5 to 0 °C)
2.5 (À5 to 0 °C)
2.5 (À5 to 5 °C)
3.0 (À5 to 5 °C)
89
80
65
94
93
0.1
0.1
200
a A: addition of LiHMDS to 5 in THF followed by bromide. B: addition
of LiHMDS to a mixture of 5 and bromide in THF. b Isolated from a
mixture of IPAc/heptane.
Scheme 5. Synthesis of bromoimidazole 12
We originally performed the dehydration/cyclization of urea 9
using a combination of Ph3P, CCl4, and TEA10 which was
demonstrated to be effective in the synthesis of a similar guanidine
derivative.4 The reaction produced 10 (as a 1:1 mixture of two
endo- and exoisomers) in 94% isolated yield. Considering that
the use of toxic CCl4 as reagent in production was not desirable,
we searched for alternative dehydration conditions. The combi-
nation of Ph3P and CCl4 has been used for the preparation of
Ph3PCl2 which is now commercially available and inexpensive,
and it is often used for N-ylide formation.11 We therefore studied
the possibility of employing this reagent for our key dehydration.
We were pleased to find that a slow addition of 1.7 equiv of
Ph3PCl2 to 9 and 3.5 equiv of TEA in acetonitrile at 20À25 °C,
gave 10 in quantitative yield (Scheme 5). Without a workup,
the resulting mixture was treated with concentrated HCl
at 40À50 °C to give crude bicyclic imidazole 11, which
contained Ph3PO, in 95% yield (HPLC assay) over two steps.
The HPLC assay yield was consistent with small-scale experi-
ments in which 11 was purified by silica gel chromatography.
Intermediate 11, crude or purified through chromatography,
was an oil that was used directly in the subsequent bromina-
tion after a simple aqueous workup of the IPAc solution12 of
crude 11.
During the pilot-plant production of 12, we had one occasion
when the bromination of crude 11 in IPAc gave only 40% of 12
on 50-kg scale due to the formation of a byproduct identified as
the ring-cleavage product 18. We found that the presence of
water in the crude IPAc solution of 11 led to the ring cleavage of
imidazole to 18. As demonstrated in Table 4, a significant
amount of 18 was found when the water content in the IPAc
solution reached 1%. For the pilot-plant incident, we concluded
that an inefficient phase cut during the aqueous workup and
insufficient removal of water by azeotropic distillation of the
IPAc solution of 11 resulted in the water levels that in the
presence of NBS cleaved the imidazole ring. Since this side
reaction was not significant when the water content was <1000
ppm, a specification was set for this particular step and this failure
mode was not observed in all subsequent batches.
The final stage of the process, namely the transformation of 12
to BIRT2584, was executed next. Br/Mg exchange of 12 was
performed by addition of iPrMgCl (1.05 equiv) in THF at
À5 °C. Upon completion of the charge of iPrMgCl, successive
addition of SO2 (1.1 equiv) as a THF solution and a suspension
of NCS (1.1 equiv) in THF to the reaction mixture at À5 °C
produced sulfonyl chloride 19. Without a workup, coupling of 19
with L-alaninamide was then carried out. A screening of conditions
1187
dx.doi.org/10.1021/op200175t |Org. Process Res. Dev. 2011, 15, 1185–1191

Organic Process Research & Development
ARTICLE
Table 3. Bromination of 11 with NBS to bromoimidazole 12
L-alaninamide and TEA or pyridine in DMF (as cosolvent) to
the reaction mixture of 19 produced BIRT2584 in only 40À50%
yield due to the formation of a large amount of sulfonic acid 20
over the prolonged reaction times required (entries 2À3).
Addition of inorganic bases such as K2CO3 under the same
conditions gave a similar result (entry 4). While the coupling
reaction using a combination of DMF and water as cosolvent and
TEA as base gave an improved yield of BIRT 2584 (entry 5), the
use of either K2CO3 or Cs2CO3 significantly reduced the
formation of 20 with a complete reaction in 6 h, resulting in
85À86% yield of BIRT 2584 after crystallization from ethanol/
water (entries 6À7). In the absence of DMF, the reaction rate
was dramatically reduced, and 24 h were required for a complete
reaction. Under these conditions more hydrolysis of the sulfonyl
chloride was also observed (entry 8).
During production in the pilot plant, we experienced a yield
fluctuation of BIRT2584 using K2CO3 as base with different
levels of 20, presumably formed due to the poor solubility of the
base in a multiphase reaction system. The search for an alter-
native base other than Cs2CO3 which gave consistent results
revealed that aq NaOH produced a result similar to those of
K2CO3 and Cs2CO3 (entry 9). It was interesting to notice that
the pH values at the end point for all reactions with >85% of
BIRT 2584 were between 9 and 10.5 which we believe to be
critical for a satisfactory reaction (such as entries 6, 7, and 9). This
observation is in agreement with the fact that both excess and
substoichiometric NaOH produced BIRT2584 in diminished
yield along with increased amounts of 20 (entries 10 and 11). In
general, the byproduct 20 was totally rejected after initial
isolation of crude BIRT2584 from a 1:5 mixture of EtOAc/
heptane. The final crystallization was performed in a 10:1 mixture
of EtOH/water to produce the desired polymorph in 92% yield.
additive
dimerization
derivatives (area %)a
isolated
yield of 12b
entry
(0.1 equiv)
time (h)
1
2c
3
4
5
6
/
11
11
13
22
2
1
63
64
58
45
83
86
/
1
PPTS
ZnBr2
TEA
K2CO3
1
0.5
1
1
1
a Area purity by HPLC. b Isolated yield by crystallization from IPA/
water. c Chromatographically pure 11 used.
Table 4. Water impact for bromination of 11
’ SUMMARY AND CONCLUSION
We developed and implemented a concise and robust process
for the production of LFA-1 inhibitor BIRT2584 on metric ton
scale. The process features an efficient synthesis of Seebach’s
template, a new methodology for the dehydration of a urea
intermediate with Ph3PCl2 and a regioselective bromination of
imidazo[1,2-a]imidazol-2-one derivatives. Furthermore, an effi-
cient one-pot process for the preparation of the sulfonamide
from a bromoimidazole using SO2 solutions was also developed
involving four in situ chemical transformations.
’ EXPERMENTAL SECTION
HPLC wt % assay HPLC wt % assay
General. All raw materials were used as received from commer-
cial sources without pretreatment. HPLC monitoring for all reac-
tions was carried out with the use of commercially available
reverse-phase columns (Zorbax Eclipse XDB-C8) eluted with
0.05% TFA (aq) and 0.05% TFA in acetonitrile. The chiral purity
of 7 was determined using a Chiralpak AD-3, 3.0 μm, 150 mm Â
3.0 mm column, eluting with 2% IPA in hexane.
(2S,5R)-3-(3,5-Dichlorophenyl)-2-isopropyl-5-methylimida-
zolidin-4-one (4). First Batch: p-Toluenesulfonic acid salt of 3
(2578 g, 5.10 mol) in EtOAc (7.0 L) was stirred with 2 N NaOH
(2.8 L, 5.60 mol) for 30 min, and the separated organic layer was
washed with water (2.0 L). Isobutyraldehyde (571 mL, 6.00 mol)
was added to this stirred EtOAc solution at 17À23 °C over a
period of 30 min. Azeotropic distillation was performed at reflux
with a DeanÀStark until no more water was collected (∼10 h).
entry scale (g) water (%)
yield of 12(%)
yield of 18(%)
1
2
3
4
5
6
100
100
100
100
100
100
0.02
0.05
0.10
0.50
1.00
2.00
93
93
90
84
69
65
<1
<1
4
8
15
18
revealed that the choices of solvents and bases had a significant
impact on the reaction, as shown in Table 5. Without a cosolvent,
the coupling of 19 with L-alaninamide in the presence of TEA was
sluggish, probably due to the poor reactivity and solubility of
L-alaninamide in THF (entry 1). Addition of a mixture of
1188
dx.doi.org/10.1021/op200175t |Org. Process Res. Dev. 2011, 15, 1185–1191

Organic Process Research & Development
ARTICLE
Table 5. One-pot synthesis of LFA-1 BIRT2584 from 12
entry
basea
cosolventb (vol % to THF)c
time (h)
yield of BIRT 2584d (wt % of 20)e
1
TEA
TEA
py
/
20
17
15
20
6
<10%
2
DMF (10)
45% (23%)
3
DMF (10)
40% (25%)
4
K2CO3
DMF (10)
51% (20%)
5
TEA
DMF (10)/H2O (10)
DMF (10)/H2O (10)
DMF (10)/H2O (10)
H2O (10)
64% (13%)
6
K2CO3
6
85% (6%)
7
Cs2CO3
6
86% (5%) (pH ≈ 9.8)f
72% (11%)
86% (4%) (pH: ∼10)f
79% (9%) (pH > 11)f
76% (10%) (pH ≈ 8)f
8
K2CO3
24
6
9
NaOH (2.4 equiv)
NaOH (2.6 equiv)
NaOH (2.2 equiv)
DMF (10)/H2O (10)
DMF (10)/H2O (10)
DMF (10)/H2O (10)
10
6
11
10
a 3 equiv for TEA, pyridine and 1.5 equiv for K2CO3 and Cs2CO3 used. b Added as L-alaninamide HCl salt solution. c Volume % to all THF used in the
first three chemical transformations. d Isolated by crystallization from EtOH/water. e Weight % assay of the reaction mixture by HPLC. f pH value at the
end point.
The mixture was then concentrated to a low volume (∼1.5 L),
and heptane (6.0 L) was added. Distillation was performed under
reduced pressure at 50À60 °C to a low volume (∼1.5 L). To the
concentrate was added heptane (6.0 L), and the solution was
stirred at 50 °C for 12 h while crystallization occurred. The
mixture was cooled to 20À25 °C and filtered, and the filter cake
was washed with heptane (1.0 L) to give 4 (1113 g, 76%) as a
light-yellow solid.
for C13H16Cl2N2O: C, 54.37; H, 5.62; N, 9.75. Found C,
54.44; H, 5.43; N, 9.63%.
(2R,5R)-3-(3,5-Dichlorophenyl)-2-isopropyl-5-methyl-1-(2,2,2-
trifluoroacetyl)imidazolidin-4-one (5). Trifluoroacetic anhydride
(400 mL, 2.83 mol) was added dropwise to a stirred solution of 4
(796 g, 2.77 mol), triethylamine (394 mL, 2.83 mol), and THF
(6000 mL) at 0 °C over a period of 90 min. The resulting solution
was allowed to reach ambient temperature and stirred for 90 min.
The mixture was concentrated under reduced pressure to a low
volume (∼1.5 L), and water (6000 mL) was added over 60 min
while seed crystals of 5 were added. The resulting slurry was
filtered to afford 1290 g (98%) of a white solid: mp 87À89 °C;
1H NMR (400 MHz, CDCl3) δ 0.74 (d, J = 7.2 Hz, 3H), 0.91 (d,
J = 6.8 Hz, 3H), 1.68 (d, J = 6.4 Hz, 3H), 2.42 (m, 1H), 4.57 (q,
J = 6.6 Hz, 1H), 6.18 (s, 1 H), 7.30 (m, 1 H), 7.45 (br s, 2 H); 13C
NMR (400 MHz, CDCl3) δ 15.8, 18.1, 20.2, 57.0, 78.0, 122.4,
127.3, 135.7, 137.4, 168.7; Anal. calcd for C15H15Cl2F3N2O2: C,
47.02; H, 3.95; N, 7.31. Found C, 47.16; H, 3.82; N, 7.24
(2R,5R)-5-(4-Trifluoromethoxybenzyl)-3-(3,5-dichlorophenyl)-
2-isopropyl-5-methyl-1-(2,2,2-trifluoroacetyl)imidazolidin-4-one
(6). To a mixture of 5 (750 g, 1.96 mol) and p-trifluoromethox-
ybenzyl bromide (513 g, 2.01 mol) in THF (2.0 L) was added a
1 M solution of LiN(TMS)2 in THF (2030 mL, 2.03 mol) at
0À5 °C over 1 h. The mixture was then stirred at this tempera-
ture for an additional 30 min and quenched with 5% NH4Cl
(3 L). A distillation was performed to remove THF followed by
addition of a 10:1 mixture of heptane and ethyl acetate (3 L). The
resulting slurry was filtered, and the cake was washed with water
(1 L). The filter cake was dried under vacuum at 40 °C to give 6
Second Batch: p-Toluenesulfonic acid salt of 3 (2578 g, 5.10
mol) in EtOAc (7.0 L) was stirred with 2 N NaOH (2.8 L, 5.60
mol) for 30 min, and the separated organic layer was washed with
water (2.0 L). Isobutyraldehyde (571 mL, 6.00 mol) was added
to this stirred EtOAc solution at 17À23 °C over a period of
30 min. Azeotropic distillation was performed at reflux with a
DeanÀStark until no more water was collected (∼10 h). After
the mixture was concentrated to a low volume (∼1.5 L) by
distillation, the filtrate from the first batch was added to the
concentrate of the second batch. Distillation was performed
under reduced pressure at 50À60 °C to a low volume (∼1.8
L). To the concentrate was added heptane (6.5 L), and the
solution was stirred at 50 °C for 12 h while crystallization
occurred. The mixture was cooled to 20À25 °C and filtered,
and the filter cake was washed with heptane (1.0 L) to give 4
1
(1391 g, 95%) as a light-yellow solid. H NMR (400 MHz,
CDCl3) δ 0.79 (d, J = 6.8 Hz, 3H), 0.99 (d, J = 6.8 Hz, 3H),
1.38 (d, J = 6.8 Hz, 3H), 2.02À1.95 (m, 2H), 3.72 (q, J = 6.1
Hz, 1H), 5.02 (d, J = 1.96 Hz, 1H), 7.17 (m, 1H), 7.41 (d, J =
1.96 Hz, 2H); 13C NMR (400 MHz, CDCl3) δ 14.1, 18.1, 18.4,
30.5, 56.2, 78.1, 120.8, 125.5, 135.4, 138.7, 174.9; Anal. calcd
1189
dx.doi.org/10.1021/op200175t |Org. Process Res. Dev. 2011, 15, 1185–1191

Organic Process Research & Development
ARTICLE
as a white solid (1016 g, 93%). 1H NMR (400 MHz, CDCl3) For
the major rotamer: δ 7.29 (d, J = 2.0 Hz, 1H), 7.10 (ABq, J = 8,0
Hz, 2H), 6.98 (ABq, J = 8,0 Hz, 2H), 6.83 (d, J = 2.0 Hz, 2H),
5.08 (s, 1H), 3.76 (d, J = 14.0 Hz, 1H), 3.12 (d, J = 14.0 Hz, 1H),
2.12À2.03 (m, 1H), 1.96 (s, 3H), 0.92 (d, J = 6.4 Hz, 3H), 0.55
(d, J = 6.4 Hz, 3H). For the minor rotamer: δ 7.26 (s, 1H), 7.10
(ABq, J = 8,0 Hz, 2H), 6.98 (ABq, J = 8,0 Hz, 2H), 6.75 (s, 2H),
5.14 (s, 1H), 3.33 (d, J = 14.4 Hz, 1H), 3.17 (d, J = 14.4 Hz, 1H),
2.30À2.22 (m, 1H), 1.92 (s, 3H), 0.79 (d, J = 6.4 Hz, 3H), 0.55
(d, J = 6.4 Hz, 3H). 13C NMR (400 MHz, CDCl3) for the major
rotamer only: δ 169.9, 155.7 (q, J = 37.0 Hz), 148.8 (d, J = 2.0
Hz), 137.3, 135.7, 133.8, 131.0, 128.1, 123.7, 121.0, 120.4 (q, J =
256.0 Hz), 115.6 (q, J = 287.0 Hz), 78.0 (q, J = 5.0 Hz), 70.1,
39.2, 36.5 (d, J = 1.0 Hz), 22.4, 20.2, 14.4; HRMS calculated for
C23H20Cl2F6N2O3 [M + H]+: 557.0827, Found: 557.0838.
p-Toluenesulfonic Acid Salt of (R)-2-Amino-3-(4-
trifluoromethoxyphenyl)-N-(3,5-dichlorophenyl)-2-methyl-
propionamide (7). A mixture of 6 (1000 g, 1.79 mol) and KOH
(111 g, 1.97 mol) in IPA (3.6 L) was heated to 60 °C for 1 h, and
then 3 M sulfuric acid (750 mL, 2.27 mol) was added. After
agitating at 60 °C for 3 h, a distillation of IPA was performed to a
low volume (∼1.5 L). IPAc (4 L) was then added followed by a
slow addition of 0.2 N KOH in water (3 L, 0.60 mol). The layers
were separated, the organic layer was distilled to a low volume
(∼1.7 L), and acetonitrle (2 L) was then charged. To this
solution was added a solution of p-toluenesulfonic acid mono-
hydrate (358 g, 1.88 mol) in acetonitrile (1.5 L), causing a slurry
of the salt to form. The product was then collected by filtration
and dried under vacuum at 40 °C to give 7 as light-yellow crystals
(902 g, 87%). 1H NMR (400 MHz, MeOD-d4) δ 7.76 (ABq, J =
8.0 Hz, 2H), 7.71 (d, J = 2.0 Hz, 2H), 7.33 (ABq, J = 8.8 Hz, 2H),
7.23À7.19 (m, 5H), 3.44 (d, J = 14.4 Hz, 1H), 3.31 (d, J = 14.4
Hz, 1H), 2.35 (s, 3H), 1.79 (s, 3H); 13C NMR (400 MHz,
MeOD-d4) δ 170.5, 150.3 (d, J = 1.0 Hz), 143.4, 141.9, 141.1,
136.2, 133.5, 133.2, 130.0, 127.0, 125.5, 122.5, 121.9 (q, J = 254.0
Hz), 120.3, 62.7, 42.8, 21.9, 21.4; HRMS calculated for
C17H15Cl2F3N2O2 [M + H]+: 407.0535, Found: 407.0538.
(R)-N-(3,5-Dichlorophenyl)-2-[3-(2,2-dimethoxyethyl)ureido]-
2-methyl-3-(4-trifluoro-methoxyphenyl)propionamide (9). To
a solution of 7 (850 g, 1.47 mol) in DMSO (1.5 L) was added 8
(365 g, 1.61 mol) in MTBE (1.5 L). After a distillation was
performed to remove MTBE, TEA (223 g, 2.21 mol) was added.
The resulting mixture was heated to 60 °C for 3 h and quenched
with addition of EtOAc (3.5 L) and water (3 L). The layers were
separated, and the organic phase was washed with water (1 L). A
distillation of EtOAc was performed to a low volume (∼1.5 L)
followed by addition of MeOH (4 L). The solution was again
distilled to a low volume (∼3.5 L), and the solids that precipitated
were collected by filtration. The filter cake was washed with MeOH
(500 mL) and dried under vacuum at 40 °C to give 9 (728 g, 92%).
1H NMR (500 MHz, DMSO-d6) δ 9.93 (s, 1H), 7.78 (d, J = 2.0 Hz,
2H), 7.28À7.22 (m, 5H) 6.21 (s, 1H), 6.13 (t, J = 6.0 Hz, 1H), 4.36
(t, J = 5.6 Hz, 1H), 3.41 (d, J = 13.2 Hz, 1H), 3.30 (s, 6H),
3.27À3.21 (m, 1H), 3.12 (d, J = 13.2 Hz, 1H), 3.12À3.05 (m, 1H),
1.26 (s, 3H); 13C NMR (500 MHz, DMSO-d6) δ 173.9, 156.8,
147.1 (d, J= 2 Hz), 141.7, 136.7, 133.8, 132.4, 122.0, 120.3, 120.1 (q,
J = 254.0), 117.9, 102.7, 59.0, 53.3, 53.2, 40.7, 23.1; HRMS
calculated for C22H24Cl2F3N3O5 [M + H]+: 538.1117, Found:
538.1127.
in acetonitrile (3 L) was slowly added a solution of Ph3PCl2
(1263 g, 3.79 mol) in acetonitrile (2 L) at 10À20 °C over 30 min.
The resulting mixture was stirred at rt for 2 h (to give 10), and
concentrated HCl (50 mL, 0.60 mol) was then added. After being
stirred at 40À50 °C for 2 h, the mixture was quenched with the
addition of K2CO3 (85 g, 0.61 mol) in water (3 L). A partial
distillation of CH3CN was next performed to a volume of about
5 L, and IPAc (4 L) was then charged. The mixture was filtered to
remove some solids (TPPO), and the layers were separated. The
organic layer was washed with water (1 L), and an azeotropic
distillation was performed until KF analysis of the solution
showed <1000 ppm water. An estimated 0.5 M solution of 11
was used for the next bromination. To the above IPAc solution of
11 was added K2CO3 (15 g, 0.11 mol) followed by a suspension
of NBS (395 g, 2.22 mol) in IPAc (2 L) at 10À20 °C over 30 min.
The resulting mixture was stirred at rt for 1 h. The mixture
was filtered, and the filtrate was distilled to a low volume (2 L).
IPA (4 L) was added, and a distillation was again performed to a
low volume (2 L, 2Â). After addition of IPA (4 L), water
(600 mL) was slowly added. The resultant slurry was stirred
for an additional 1 h, and the product was collected by filtration.
The filter cake was washed with IPA (500 mL), and dried under
vacuum at 40 °C to give 12 as a light-yellow solid (990 g, 83%).
1H NMR (400 MHz, CDCl3) δ 7.57 (d, J = 2.0 Hz, 1H), 7.28 (m,
1H), 7.03 (ABq, J = 8.4 Hz, 2H), 6.97 (ABq, J = 8.4 Hz, 2H), 6.87
(s, 1H), 3.55 (d, J = 14.0 Hz, 1H), 3.35 (d, J = 14.0 Hz, 1H), 1.94
(s, 3H). 13C NMR (400 MHz, CDCl3) δ 173.8, 149.9 (d, J = 1.0
Hz), 146.4, 135.4, 134.1, 131.8, 130.9, 128.8, 127.4, 121.0, 120.4,
120.3 (q, J = 256.0 Hz), 95.2, 68.3, 41.9, 22.0; HRMS calculated
for C20H13BrCl2F3N3O2 [M + H]+: 533.9593, Found: 533.9599.
A small sample of 10 was isolated as an oil (a 1:1 mixture of
1
endo- and exo-isomers) by chromatography on silica gel. H
NMR (400 MHz, CDCl3) δ 7.90 (s, 1H), 7.73 (d, J = 8.4 Hz,
1H), 7.51 (d, J = 8.0 Hz, 1H), 7.44 (m, 2H), 6.80 (m, 1H), 4.51
(s, 2H), 4.17 (q, J = 7.2 Hz, 2H), 1.21 (t, J = 7.2 Hz, 3H); 13C
NMR (400 MHz, CDCl3) δ 191.9 (d, J = 2.0 Hz), 167.5, 162.4
(d, J = 249.0 Hz), 161.1, 140.3 (d, J = 7.0 Hz), 132.7 (dq, J = 7.5,
33.6 Hz), 132.6, 130.1, 129.0, 127.4, 122.9 (q, J = 272.0 Hz),
122.2 (m), 120.1 (d, J = 22.0 Hz), 117.0 (m), 113.7, 65.5, 61.6,
14.0; HRMS calculated for C22H22Cl2F3N3O4 [M + H]+:
520.1012, Found: 520.1034.
A small sample of 11 was isolated as an oily residue by
1
chromatography on silica gel. H NMR (400 MHz, CDCl3)
δ 7.74 (d, J = 2.0 Hz, 2H), 7.26 (t, J = 1.6 Hz, 1H), 7.03 (ABq, J =
8.0 Hz, 2H), 6.98 (d, J = 2.0 Hz, 1H), 6.93 (d, J = 2.0 Hz, 1H),
6.92 (ABq, J = 8.0 Hz, 2H), 3.33 (d, J = 14.0 Hz, 1H), 3.21 (d, J =
14.0 Hz, 1H), 1.79 (s, 3H); 13C NMR (400 MHz, CDCl3)
δ 174.9, 148.8 (d, J = 2.0 Hz), 145.8, 135.3, 134.8, 131.8, 131.0,
129.2, 127.0, 120.9, 120.3 (q, J = 256.0 Hz), 120.1, 111.1, 66.1,
43.9, 23.1; HRMS calculated for C20H14Cl2F3N3O2 [M + H]+:
456.0487, Found: 456.0508.
(S)-2-[(R)-7-(3,5-Dicholrophenyl)-5-methyl-6-oxo-5-(4-triflu-
oromethoxybenzyl)-6,7-dihydro-5H-imidazo[1,2-a]imidazole-
3-sulfonylamino]propionamide (BIRT2584). To a solution of
12 (1000 g, 1.87 mol) in THF (2.0 L) was added a 2.0 M solution
of iPrMgCl (1968 mL, 1.97 mol) at À15 to À5 °C over 45 min.
After agitating at this temperature for an additional 10 min, a
prepared SO2 solution (133 g, 2.06 mol) in THF (400 mL) was
added at À15 to À5 °C over 45 min, followed immediately by
the addition of a suspension of NCS (292 g, 2.19 mol) in
THF (1200 mL) over 30 min. The resultant mixture was aged
at À5 °C for 10 min, then NaOH (176 g, 4.40 mol) was charged
(R)-3-(4-Trifluoromethoxybenzyl)-1-(3,5-dichlorophenyl)-5-
bromo-3-methyl-1H-imidazo[1,2-a] imidazole-2-one (12). To
a suspension of 9 (1200 g, 2.23 mol) and TEA (790 g, 7.81 mol)
1190
dx.doi.org/10.1021/op200175t |Org. Process Res. Dev. 2011, 15, 1185–1191

Organic Process Research & Development
ARTICLE
to the reaction mixture followed by the addition of L-alaninamide
HCl salt (302 g, 2.42 mol) in DMF (550 mL) and water (550 mL)
over 15 min, keeping the internal temperature À20 °C. After
having stirred at 18 to 23 °C for 6 h, the mixture was quenched
with water (4 L) and ethyl acetate (2.4 L). The layers were
separated, and the organic layer was distilled to a low volume
(∼2.3 L), and EtOAc (5.0 L) was then added. The solution
was then washed with water (3.0 L) and distilled to a low volume
(2.0 L). A slow addition of heptane (6.0 L), antisolvent to the
organic solution, was then made, and the resulting slurry was
stirred at rt for 2 h. The crude product was collected by filtration
to give BIRT 2584 as white crystals (975 g, 86%) after drying. 1H
NMR (400 MHz, DMSO-d6) δ 8.0 (s, 1H), 7.63 (t, J = 1.6 Hz,
1H), 7.46 (d, J = 2.0 Hz, 2H), 7.44 (s, 1H), 7.42 (s, 1H), 7.16
(ABq, J = 8.0 Hz, 2H), 7.14 (s, 1H), 6.99 (ABq, J = 8.0 Hz, 2H),
3.84 (q, J = 7.2 Hz, 1H), 3.77 (d, J = 14.0 Hz, 1H), 3.28 (d, J =
14.0 Hz, 1H), 1.96 (s, 3H), 1.22 (d, J = 7.2 Hz, 3H); 13C NMR
(400 MHz, MeOD-d4) δ 177.1, 175.4, 150.3, 150.1 (d, J = 2.0
Hz), 136.5, 135.9, 135.3, 134.4, 132.3, 128.7, 128.3, 122.6, 122.0,
121.8 (q, J = 254.0 Hz), 71.6, 53.5, 43.4, 22.2, 19.6; HRMS
calculated for C23H20Cl2F3N5O5S [M + H]+: 606.0587, Found:
606.0590.
(10) Appel, R.; Kleinstuck, R.; Ziehn, K. Chem. Ber. 1971, 104, 1335.
(11) (a) Wamhoff, H.; Schupp, W.; Kirfel, A.; Will, G. J. Org. Chem.
1986, 51, 149. (b) Wamhoff, H.; Haffmanns, G.; Schmidt, H. Chem. Ber.
1983, 116, 1691.
(12) Amounts of 11 in many IPAc solution batches, consistently
corresponded to about 95% yield from 9 over two steps, as measured by
weight % HPLC assay. Halogenation reagents were therefore charged
based on this value.
(13) For examples, see: (a) Choshi, T.; Yamada, S.; Sugino, E.;
Kuwada, T.; Hibino, S. J. Org. Chem. 1995, 60, 5899. (b) Matthews,
H. R.; Rapoport, H. J. Am. Chem. Soc. 1973, 95, 2297. (c) Takeuchi, Y.;
Yeh, H. J. C.; Kirk, K. L.; Cohen, L. A. J. Org. Chem. 1978, 43, 3565.
(14) For examples, see: (a) Jensen, M. S.; Hoerrner, R. S.; Li, W.;
Nelson, D. P.; Javadi, G. J.; Dormer, P. G.; Cai, D.; Larsen, R. D. J. Org.
Chem. 2005, 70, 6034. (b) Rival, Y.; Grassy, G.; Michel, G. Chem. Pharm.
Bull. 1992, 40, 1170.
(15) For examplex, see: (a) Hua, D. H.; Zhang, F.; Chen, J. J. Org.
Chem. 1994, 59, 5084–5087. (b) Kn€olker, H.-J.; Hitzemann, R. Tetra-
hedron Lett. 1994, 35, 2157.
(16) On the basis of MS data, they are possibly hydrates of imidazole
dimerization. The ring strain associated with these dimers may lead to
the addition of water.
(17) (a) Monguchi, D.; Yamamura, A.; Fujiwara, T.; Somete, T.;
Mori, A. Tetrahedron Lett. 2010, 51, 850. (b) Stachel, S. J.; Habeeb, R. L.;
Van Vranken, D. L. J. Am. Chem. Soc. 1996, 118, 1225.
(18) Dunstan, S.; Henbest, H. B. J. Chem. Soc. 1957, 4905.
(19) March, J. Advanced Organic Chemistry, 4th ed.; Wiley-Interscience:
Hoboken, NJ, 1992; pp 613À615.
’ ASSOCIATED CONTENT
Supporting Information. Copies of 1H/13C NMR spec-
S
b
tra for 4À7, 9À14, 18, 20, BIRT2584. This material is available
free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*xiao-jun.wang@boehringer-ingelheim.com.
’ REFERENCES
(1) (a) Wu, J.-P.; Emeigh, J.; Gao, D. A.; Goldberg, D. R.; Kuzmich,
D.; Miao, C.; Potocki, I.; Qian, K. C.; Sorcek, R. J.; Jeanfavre, D. D.;
Kishimoto, K.; Mainolfi, E. A.; Nabozny, G., Jr.; Peng, C.; Reilly, P.;
Rothlein, R.; Sellati, R. H.; Woska, J. R., Jr.; Chen, S.; Gunn, J. A.;
O’Brien, D.; Norris, S. H.; Kelly, T. A. J. Med. Chem. 2004, 47, 5356.
(b) Last-Barney, K.; Davidson, W.; Cardozo, M.; Frye, L. L.; Grygon,
C. A.; Hopkins, J. L.; Jeanfavre, D. D.; Pav, S.; Qian, C.; Stevnson, J. M.;
Tong, L.; Zindell, R.; Kelly, T. A. J. Am. Chem. Soc. 2001, 123, 5643.
(c) Kelly, T. A.; Jeanfavre, D. D.; McNeil, D. W.; Woska, J. R., Jr.; Reilly,
P. L.; Mainolfi, E. A.; Kishimoto, K. M.; Nabozny, G. H., Jr.; Zinter, R.;
Bormann, B.-J.; Rothlein, R. J. Immunology 1999, 163, 5173.
(2) (a) Seebach, D.; Sting, A. R.; Hoffmann, M. Angew. Chem., Int.
Ed. Engl. 1996, 35, 2708. (b) Yee, N. K. Org. Lett. 2000, 2, 2781.
(3) Wang, X.-j.; Zhang, L.; Xu, Y.; Krishnamurthy, D.; Varsolona, R.;
Nummy, L.; Shen, S.; Frutos, R. P.; Byrne, D.; Chung, J. C.; Farina, V.;
Senanayake, C. H. Tetrahedron Lett. 2005, 46, 273.
(4) Frutos, R. P.; Johnson, M. Tetrahedron Lett. 2003, 44, 6509.
(5) Frutos, R. P.; Eriksson, M.; Wang, X.-j.; Byrne, D.; Varsolona, R.;
Johnson, M. D.; Nummy, L.; Krishnamurthy, D.; Senanayake, C. H. Org.
Process Res. Dev. 2005, 9, 137.
(6) Wang, X.-j.; Xu, Y.; Zhang, L.; Krishnamurthy, D.; Wirth, T.;
Nicola, T.; Senanayake, C. H. Org. Lett. 2010, 12, 4412.
(7) Frutos, R. P.; Stehle, S.; Nummy, L.; Yee, N. Tetrahedron:
Asymmetry 2001, 12, 101.
(8) Carbamate 8 was prepared by the reaction between aminoace-
tyaldehyde dimethyl acetal and phenyl chloroformate in the presence of
TEA in MTBE. After filtration to remove TEA HCl salt, the concen-
trated solution (can be stored) was used directly.
(9) Thavonekham, B. Synthesis 1997, 1189.
1191
dx.doi.org/10.1021/op200175t |Org. Process Res. Dev. 2011, 15, 1185–1191