Development of an Efficient Manufacturing Process for a Key Intermediate in the Synthesis of Edoxaban

Article abstract of DOI:10.1021/acs.oprd.8b00413

We report the development of a novel synthetic method to access a key intermediate in the synthesis of edoxaban. The main features of the new synthetic method are an improvement in the approach for the synthesis of a key chiral bromolactone, application of an interesting cyclization reaction utilizing neighboring group participation to construct a differentially protected 1,2-cis-diamine, and implementation of plug-flow reactor technology to enable the reaction of an unstable intermediate on multihundred kilogram scale. The overall yield for the preparation of edoxaban was significantly increased by implementing these changes and led to a more efficient and environmentally friendly manufacturing process.

Full text of DOI:10.1021/acs.oprd.8b00413

Article
pubs.acs.org/OPRD
Cite This: Org. Process Res. Dev. XXXX, XXX, XXX−XXX
Development of an Efficient Manufacturing Process for a Key
Intermediate in the Synthesis of Edoxaban
Makoto Michida,*,† Hideaki Ishikawa,† Takeshi Kaneda,† Shinya Tatekabe,‡ and Yoshitaka Nakamura§
†Process Technology Research Laboratories (PTRL), Daiichi Sankyo Co., Ltd., 1-12-1 Shinomiya, Hiratsuka-shi, Kanagawa
254-0014, Japan
‡Plant Management Department, Daiichi Sankyo Chemical Pharma Co., Ltd., 477 Takada, Odawara-shi, Kanagawa 250-0216, Japan
§Global Supply Chain - Technology Function, Daiichi Sankyo, Inc., 211 Mt. Airy Road, Basking Ridge, New Jersey 07920, United
States
S
* Supporting Information
ABSTRACT: We report the development of a novel synthetic method to access a key intermediate in the synthesis of
edoxaban. The main features of the new synthetic method are an improvement in the approach for the synthesis of a key chiral
bromolactone, application of an interesting cyclization reaction utilizing neighboring group participation to construct a
differentially protected 1,2-cis-diamine, and implementation of plug-flow reactor technology to enable the reaction of an
unstable intermediate on multihundred kilogram scale. The overall yield for the preparation of edoxaban was significantly
increased by implementing these changes and led to a more efficient and environmentally friendly manufacturing process.
KEYWORDS: process development, enzymatic resolution, flow reactor, 1,2-cis-diamine, rearrangement reaction, aziridine
As shown in Figure 1, Edoxaban (1) can be accessed via the
combination of three structurally distinct units: thiazole
carboxylic acid 2, chiral cyclohexane cis-diamine 3, and oxalate
4.4 Notably, 3 features three asymmetric centers on the
cyclohexane ring. When designing a synthetic route for 1, our
most important consideration was how to efficiently control
the stereocenters of 3 while simultaneously installing the
differentially protected vicinal cis-diamine moiety. Herein, we
describe the establishment of a highly efficient manufacturing
procedure of key intermediate 3 which was identified by
exploring alternative synthetic approaches to 3 from
inexpensive and commercially available starting materials.
INTRODUCTION
■
Several new direct oral anticoagulants (DOACs) have been
approved for the prevention of stroke in patients with atrial
fibrillation (AF) and for the treatment and prevention of
venous thromboembolism (VTE) recurrence.1 DOACs inhibit
Factor Xa (FXa) in the coagulation cascade and represent
clinical alternatives to traditional vitamin K antagonists such as
warfarin, which possess limitations such as risk of bleeding,
strong drug−drug interactions, and the requirement of
frequent monitoring.2 Edoxaban (1) is a once-daily DOAC
that has been launched under the trade names Lixiana and
Savaysa for the prevention of AF and VTE.3
RESULTS AND DISCUSSION
■
The original manufacturing route for the preparation of 3 is
shown in Scheme 1.5 This process started with a classical
resolution of rac-5 with (R)-phenylethylamine (PEA) to afford
the (S)-5/PEA salt.6 This was followed by a bromolactoniza-
tion to give the corresponding optically pure lactone 6 (25%
isolated yield from rac-5). Then the lactone moiety in 6 was
opened by dimethylamine (Me2NH) to form bromohydrin 7,
which was converted to amino alcohol 9, with excellent
regioselectivity, via the corresponding epoxide 8 following
treatment with aqueous ammonia. Boc protection of the
resultant amino group gave 10, and subsequent mesylation of
the secondary alcohol afforded 11 as a crystalline solid (63%
isolated overall yield from 6). Construction of the cis-diamine
functionality was achieved by azidation followed by hydro-
genation. Although approximately 10% of undesired trans-12
Special Issue: Japanese Society for Process Chemistry
Received: November 30, 2018
Figure 1. Synthesis of edoxaban (1).
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Original Synthetic Route To Prepare cis-Diamine 3
was generated under the optimized reaction conditions, trans-
12 was converted to the corresponding trans-3 and was
effectively removed during the crystallization and isolation of
cis-3. As a result, pure cis-diamine 3 was obtained in 64% yield
from 11.
This process is reproducible and was capable of producing 3
at production scale; however, a critical evaluation of the
process indicated that several aspects required improvement
due to unsatisfactory processes. Especially, identifying a highly
stereoselective synthesis of the cis-diamine moiety was the
most important issue from the viewpoint of quality control and
cost reduction. In order to solve those problems, we focused
on finding alternative routes for the isolable intermediates such
as bromolactone 6 and ultimately diamine 3.
Figure 2. Alternative synthetic approach to 6.
Scheme 2. Bromolactonization of rac-5 in Water
Preparation of Bromolactone 6. In the original process,
(S)-5 is separated from rac-5 by diastereomeric salt
crystallization as the (S)-5/(R)-PEA salt. However, to obtain
the desired optical purity, repeated recrystallizations of the
initially isolated salt (5−6 times) were required to attain >98%
ee of bromolactone 6. Screening of additional chiral amines
failed to identify an improvement in the optical purity of the
corresponding diastereomeric salt.
Since the resolution of rac-5 was unavoidable, we examined
enzymatic resolution of racemic 6 as an alternative approach
(Figure 2). Usually, it is preferable to conduct an optical or
enzymatic resolution step as early as possible in the synthetic
route to avoid carrying along the undesired isomer. However,
since rac-6 is easily prepared from inexpensive rac-5 and a
brominating reagent, we focused our efforts on the resolution
of rac-6 rather than rac-5.
versus 80%), and we speculated that this was due to the
kinetically formed bromonium intermediate in an approx-
imately 1:1 ratio. Fortunately, the undesired bromonium
intermediate was hydrolyzed prior to isomerization when the
reaction was conducted in water. Although the isolated yield
was lower than the original process, this method appeared to
be superior in terms of cost and productivity.
In the original process, bromolactonization of 5 was carried
out in dichloromethane followed by a solvent exchange to
water for crystallization. To avoid the use of organic solvents
and reduce the number of operations, we investigated the
bromolactonization in water (Scheme 2). Under these reaction
conditions, it was possible to obtain crystalline rac-6 directly
from the reaction mixture in moderate yield with high purity.
The isolated yield was lower than the original method (50%
We subsequently explored enzymatic kinetic resolution of
rac-6 as the substrate via the screening of multiple enzymes
(Table 1).7 Using lipases and proteases, poor reactivity and
selectivity were attained (entries 1−7). On the other hand, an
esterase showed excellent reactivity and selectivity (entry 8).
In this reaction system, the corresponding hydrolyzed
carboxylic acid 13 was soluble in water, and the desired
bromolactone 6 was insoluble, facilitating isolation and
B
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Implementation of these improvements resulted in the
replacement of a stoichiometric amount of (R)-PEA with a
catalytic amount of esterase, decreased the total number of
isolations by filtration from 7 to 3, and reduced the amount of
organic solvents used by 90%. Although the overall yield and
the purity were almost the same, the material cost of 6 was
decreased by 50% when compared to the original method.
Alternative Synthetic Route Exploration of 3. Another
issue in the original route was the introduction of an amino
group into the 4-position of the cyclohexane ring to form the
vicinal cis-diamine. In the original process, the use of sodium
azide, which has a potential to form explosive hydrazoic acid,8
and the use of a transition metal hydrogenation catalyst to
produce the 4-amino group were required. Furthermore, even
under optimized conditions, more than 10% of the undesired
trans-stereoisomer was generated during the displacement of
the mesylate. Despite our efforts, the amount of the undesired
stereoisomer could not be reduced. While significant progress
was made on the preparation of 6, the synthetic route to 3 still
had several processes that required improvement in terms of
productivity, cost of goods, and the use of hazardous reagents.
Before exploring an alternative synthetic route, we
considered the potential causes of the low stereoselectivity in
the azidation reaction used to convert 11 to cis-12. The
generation of trans-12 can be rationalized via neighboring
group participation of the Boc group9 as shown in Scheme 4.
Table 1. Screening of Enzymatic Resolution of rac-6
entry
enzymes
strain
ee of 6
1
2
3
4
5
6
7
8
9
Lipase 1
Lipase 2
Lipase 3
Lipase 4
Protease 1
Protease 2
Protease 3
Esterase 1
Esterase 2
Burkholderia sp.
Pseudomonas sp.
Candida sp.
Rhizopus sp.
Aspergillus sp.
Bacillus sp.
Bacillus sp.
E. coli (Recombinant)
E. coli (Recombinant)
0%
0%
0%
0%
−48%
−5%
4%
100%
−100%
purification. After the completion of the reaction, the desired
stereoisomer 6 was obtained by filtration.
Further optimization of the reaction conditions included the
investigation of several parameters such as
• the amount of esterase
• pH
• temperature
The reactivity of esterase-1 was influenced by the pH of the
reaction media and was readily deactivated under acidic
conditions (pH < 5). Since the pH decreased as the reaction
progressed owing to the coproduction of carboxylic acid 13,
control of the pH with a pH-stat utilizing a K2CO3 aqueous
solution for neutralization was implemented. The acceptable
range of pH for esterase-1 was determined to be from 6 to 8,
and considering the instability of the lactone toward hydrolysis,
the target pH value was set to 7 (see Supporting Information).
The progress of the reaction was monitored by examining the
optical purity of the remaining solids, and when the optical
purity reached 95% ee, the addition of the neutralizing K2CO3
solution was stopped. In the isolated crude 6, which was
obtained in 49% yield, there remained a small amount of
insoluble enzymatic protein derived from esterase-1. To
remove this entrained residual protein, crude 6 was dissolved
in acetone and filtered which completely removed the residual
insoluble protein. Water was added to the acetone solution of
6 to crystallize 6 which was isolated by filtration in 92%
isolated yield with 98% ee.
Scheme 4. Proposed Mechanism for the Formation of trans-
12
The improved manufacturing method for the preparation of
optically pure bromolactone 6 is summarized in Scheme 3.
We postulated that trans-12 can be formed via initial
displacement of the mesylate in 11 to form the stabilized
carbocation 14, which can then undergo subsequent displace-
ment with azide to form the undesired trans-12.
Scheme 3. New Approach to 6
Inspired by this side reaction, we wondered if we could
utilize the neighboring group effect to control the stereo-
chemistry during the synthesis of the cis-diamine. Ultimately,
we decided on an approach that utilizes an N-protecting group
having a tether (structure A) with a terminal nitrogen as a
nucleophile, which, in turn, could be used to displace the
neighboring mesylate by an intramolecular cyclization
(structure B). Finally, removal of the tether would give the
desired cis-diamine (structure C). An overview of our proposed
strategy is presented in Figure 3.
As shown in Scheme 5, we designed and prepared substrate
17 for the proposed cyclization which has a sulfonyl group as a
tether. A Burgess-type reagent 15 was used for chemoselective
C
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 3. Synthetic strategy for cis-diamine.
desired product 18 was not observed (entry 1). Instead, the
major product was tentatively assigned to be aziridine 19 based
on the LC-MS and NMR data obtained on the crude reaction
mixture, which was not isolable due to the expected low
stability of this strained ring system. Surprisingly, a small
amount of regioisomer 20 was also obtained.10 The amount of
20 increased as the reaction progressed, even after 17 had been
completely consumed. This result likely indicates that this
reaction proceeded via initial formation of 19 as a reactive
intermediate, which is converted to 20. In fact, when the
reaction time was extended to 8 h, the aziridine 19 was
completely consumed, and converged to 20 (entry 2).
Based on our observations, we assumed that the reaction
proceeds through an intramolecular four-step sequence as
shown in Figure 4. Namely:
Scheme 5. Preparation of 17
sulfonylation of the aminoalcohol 9. Burgess-type reagents are
generally unstable in water; however, due to the solubility of 9
and the use of an aqueous ammonia solution in the previous
step, we were unable to avoid performing this reaction in the
presence of water. We optimized the reaction conditions by
using an aqueous solution of 9. It was found that the
sulfonylation reaction proceeded much faster than decom-
position of 15, and the desired sulfonamide 16 was obtained
exclusively in high yield with excellent chemoselectivity.
Mesylation of the hydroxyl group was carried out to obtain
substrate 17 for use in the subsequent cyclization reaction.
Our initial attempts at the cyclization of 17 are shown in
Table 2. In the presence of triethylamine at 80 °C for 1 h, the
(1) cyclization occurred to form aziridine 19
(2) aziridine ring opening by the pendant dimethylamide
group
(3) proton transfer
(4) recyclization to form 20
All steps in this proposed reaction sequence occurred
through intramolecular reactions to form the apparently
rearranged product with excellent stereoselectivity.
Further support for this proposed mechanism was obtained
when an ethyl ester 21 was used under the cyclization
conditions, and the corresponding aziridine 23 was also
formed (Scheme 6).11 However, 23 remained unchanged and
cyclic sulfamides 22 and 24 were not observed which supports
our proposal of the participation of the dimethylamide group
in the formation of 20. In addition, 23 possessed sufficient
stability to be isolated by silica gel column chromatography.
This result indicates that the dimethylamide group is essential
for the intramolecular cleavage of the aziridine in 19 which
ultimately leads to the formation of the desired cis-diamine
functionality.
Table 2. Cyclization of 17
HPLC area%
entry
reaction time
18
19
20
As we mentioned earlier, we focused this investigation on
finding alternative routes for the isolable intermediates,
including intermediate 3, since this approach would minimize
the impact on the downstream process and limit the need for
1
2
1 h
8 h
not detected
not detected
78%
not detected
9%
71%
Figure 4. Plausible mechanism for the formation of 20.
D
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 6. Cyclization of Ethyl Ester 21
Scheme 7. Cyclization of Boc-Protected Substrate
change control. Considering these unexpected results, we
decided to employ the Boc-protected substrate 27 instead of
continuing to explore the use of Cbz-protected 17 to
determine if this new approach could be used to our advantage
to prepare the intermediate 3 directly. This would produce 3 in
fewer steps because the Boc group would be installed via the
Boc-protected Burgess type reagent 25 which can be easily
prepared by using t-BuOH in situ. Based on our working
hypothesis, the Boc-protected substrate 27 was prepared from
9 in good yield in two steps, and the key reaction proceeded
smoothly to produce 28 in 86% yield with excellent regio- and
stereoselectivity (Scheme 7).
amount of pyridine and water with the goal of telescoping the
reaction from the previous step (entry 4). After optimization, it
was shown that the four-step cyclization sequence and
desulfonylation reactions (27 to 3) could be conducted in
one pot and 3 was isolated in 85% yield as the oxalate salt with
>99% purity.
The revised synthetic route to 3 is shown in Scheme 8. The
new optimized reaction sequence resulted in a substantial
increase in the overall yield (57% versus 40% from 6) and a
significant reduction in the cost of goods when compared to
the initial route without any disadvantages such as increase in
impurities, reaction steps, number of isolation, or the use of
hazardous reagents.
Application of Plug-Flow Reactor for Scale-up. In the
course of the scale-up studies, a decrease in the yield of 27 was
observed (Table 4). We investigated the reason for the low
yield and confirmed that the intermediate 30 was not stable
under the reaction conditions (Figure 5). Namely, chlor-
osulfonyl isocyanate (CSI) reacted with t-BuOH to form
intermediate 30 having the required Boc group; however, 30 is
unstable under acidic conditions and partially decomposed to
sulfamoyl chloride and isobutene. Since the addition reaction is
highly exothermic, it resulted in an extended reaction time as
the scale increased in order to maintain control of the
temperature. Furthermore, isobutene generated due to
decomposition of 30 reacted with CSI to form reactive
byproducts which consumed the amino alcohol 9 in the
sulfonylation step.13 A combination of all these side reactions
resulted in an increase in the amount of reactive byproducts
and a decrease in the yield of 27 upon scale-up as a result of
the extended reaction time to prepare 30.
With an expeditious route to 28 in hand, we turned our
attention to the development of the desulfonylation of 28
(Table 3) to produce 3. Recently, desulfonylation of cyclic
Table 3. Desulfonylation of 28
volume vs 28
HPLC area%
entry
pyridine
H2O
CH3CN
28
3
29
1
2
3
4
1
5
5
1
5
5
1
0.5
12.5
N.D.
0.1
0.25
40.6
96.3
96.5
62.5
13.5
0.4
none
5
0.4
0.3
To suppress the degradation reactions, it was imperative to
add CSI rapidly while keeping the reaction temperature low
and to reduce the residence time of 30. However, in the case of
multihundred kilogram scale production, the cooling capacity
was not sufficient for the needed control of the reaction
temperature in a batch mode. Therefore, we sought an
alternative approach: utilizing a plug-flow reactor that could
offer several advantages for the process from the viewpoint of
large scale manufacturing.14 The use of a plug-flow reactor
would not only offer highly efficient heat exchange and mixing,
sulfonamides has been reported using pyridine in the presence
of water or alcohols for the synthesis of diamines.12 In the
course of the optimization study for the reaction, it was found
that the chemoselectivity between the Boc carbonyl group and
the sulfonyl group depended on the ratio of pyridine and
water. Interestingly, when the amount of water was decreased,
the amount of the Boc-deprotected byproduct 29 also
decreased (entries 1−3). Further optimization was achieved
by performing the desulfonylation in CH3CN, with a reduced
E
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 8. Improved Synthetic Route of 3 from 6
mixer at constant flow rates with syringe pumps, and then
intermediate 30 was passed through a tube in an ice bath into a
Et3N/CH3CN solution to produce 25. Residence time could
be adjusted easily by varying the tube length and the flow rates.
The effect of the flow system was evaluated by comparing the
results to the reaction performed in a batch reactor (Table 5).
In the case of the conventional batch mode, the yield of 26
decreased when the reaction time of Step-1 was extended
(entries 1 and 2). On the other hand, the flow system provided
high yields by shortening the residence time of 30 (entries 3
and 4). Even when the total reaction time (Step-1 + Step-2)
increased, the amount of impurities were significantly reduced
(entry 4) in the flow system. These results indicate that the
application of a plug-flow reaction system can be an effective
alternate for scale-up.
Table 4. Yields of 27 in Scale-up Study
scale
yield of 27
10 g
100 g
1 kg
70%
67%
66%
63%
100 kg
Prior to implementation on large-scale we performed a scale-
up study. The key point of the process design was to control
the decomposition of the unstable intermediate 30 and
produce 25 in an acceptable yield. However, to determine
the appropriate conditions for the continuous plug-flow
reactor, not only the various process parameters such as flow
rate, temperature, and reagent concentration needed to be
considered, but also equipment specifications such as the
piping, heat exchanger, and mixer.
We determined these interactive parameters by process
modeling and simulation. The conceptual flowchart of our
approach is shown in Figure 7. The criterion of the yield for
production of 30 was set to more than 99% since circum-
venting the decomposition was critical to the yield and quality
of 26. On the other hand, the yield of 25 was set to more than
95% because the decomposition of 25 did not have a critical
impact on the quality of 26. In addition, from an engineering
point of view, the total operating time for preparation of 25
was set to 10−15 h to consider the limitation of the cooling
capacity in the temperature control system and to ensure an
efficient time schedule to enable sequential commercial
manufacturing.
Figure 5. Side reactions in preparation of 25.
but the residence time of unstable reaction intermediates such
as 30 could be significantly reduced. To obtain a proof-of-
concept for this approach, we conducted a preliminary
investigation of the process using simple equipment (Figure
6). CSI and t-BuOH/CH3CN solutions were supplied to a T-
The schematic diagram of process modeling used for
predicting the correlations of the yield of 30 and 25 with the
operating time is shown in Figure 8. Commercially available
and inexpensive equipment, such as general piping, an inline
static mixer, a double pipe heat exchanger, and a standard
batch reactor, were selected to minimize the required
investment in the manufacturing plant if this process was
implemented. The overall process could be divided into three
steps:
(i) flow reaction step
(ii) cooling step
(iii) batch reaction step
Figure 6. Experimental apparatus for preliminary investigation.
F
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 5. Comparison of Experimental Results between Batch Mode and Flow Mode
addition time
entry
mode
Step-1
Step-2
reaction yield of 26 from 9
1
2
3
4
batch
batch
flow
1 h
5 h
17 s
1.5 h
1.5 h
1.0 h
10 h
91%
85%
94%
93%
a
a
flow
17 s
a
Residence time of Step-1.
(2) Reaction heats for the preparation of 30 and 25 were
measured respectively in batch mode with a reaction
calorimeter.
(3) Physical properties required for process modeling such
as density, viscosity, and specific heat were measured.
(4) Process modeling including reaction rate analysis was
performed by using process data obtained above, and a
model was developed which was capable of calculating
the predicted yields of 30 and 25 and the operating time
by input of the operating conditions and equipment
specifications.
(5) Model validation was conducted. Namely, the appropri-
ateness was confirmed by the comparison of the
representative experimental results in flow mode and
the simulation results.
Figure 7. Conceptual flowchart for determination of process.
(6) Case studies were conducted under each condition. The
effect of operating conditions and equipment specifica-
tions on process results was investigated in a systematic
manner. Consequently, we were able to determine what
conditions would lead to a robust process for large-scale
manufacturing.
The process parameters in each step could be derived
respectively by well-known conventional equations. The yield
of 30 and 25 and the operating time could be quantitatively
predicted by simultaneous differential equations describing the
reaction rate and the heat balance. For the determination of
process parameters, we did the following:
(1) Decomposition rates for 30 and 25 were measured
respectively in batch mode under several temperature
conditions.
In addition, since numerous literature references indicated
that the effect of mixing would be quite important especially in
the case of instantaneous reactions,15 we evaluated the ratio of
Figure 8. Schematic diagram for process modeling.
G
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 6. Case Study of Manufacturing Conditions and Results with Scale-up Simulation
a
Case
1
2
3
4
5
6
7
CSI flow rate
0.73 L/min
0 °C
0.18 wt.fr
4.87 L/min
78 °C
0.74 L/min
0 °C
0.18 wt.fr
4.98 L/min
78 °C
0.62 L/min
0 °C
0.13 wt.fr
5.53 L/min
62 °C
0.62 L/min
−15 °C
0.13 wt.fr
5.43 L/min
50 °C
0.74 L/min
−15 °C
0.13 wt.fr
6.55 L/min
50 °C
0.46 L/min
−15 °C
0.10 wt.fr
5.50 L/min
35 °C
0.57 L/min
−15 °C
0.10 wt.fr
6.87 L/min
35 °C
t-BuOH/CH3CN temperature
t-BuOH/CH3CN concentration
t-BuOH/CH3CN flow rate
maximum temperature of 30 reaction
Heat exchanger
no
install
no
no
install
no
no
temperature of 25 reaction
0 °C
0 °C
0 °C
0 °C
0 °C
0 °C
5 °C
operating time: criteria ≥10 h, ≤ 15 h
30 yield: criteria ≥99%
25 yield: criteria ≥95%
10.2 h
81.40%
96.20%
10.0 h
97.10%
96.30%
12.1 h
96.10%
95.60%
12.1 h
99.10%
95.60%
10.0 h
99.90%
96.30%
16.2 h
99.80%
94.10%
13.0 h
99.80%
93.70%
a
Other conditions: CSI temperature 10 °C, brine temperature −20 °C, piping diameter 15 mm, piping length 20 m.
Scheme 9. Summary of Improvements
the reaction rates between the main reaction and the
decomposition of 30. As a result, it was found that the
degradation rate of 26 was much slower (400 000 times) than
the main reaction. The results implied that the reaction would
not be mixing-sensitive.16 To confirm this possibility, we
conducted several experiments with changing hydrodynamics
(see Supporting Information). Since the differences in the yield
or purity were not observed across all the conditions evaluated,
we concluded that a general plug-flow system would be
sufficient for scale-up
The summary of results from the case studies with
simulation are shown in Table 6. In Case Study 1, the reaction
conditions employed were based on the current batch
conditions and provided 30 in low yield. By installing a heat
exchanger (Case Study 2), it was expected that the yield of 30
would be increased by suppression of the decomposition of 30
in the piping. However, it was found that optimization of the
reaction conditions was more effective than the installation of a
heat exchanger. Namely, increasing the amount of CH3CN
used for the dilution of t-BuOH effectively reduced the
maximum temperature of a synthesis reaction (MTSR, Case
Study 3). In addition, a lower initial temperature of the t-
BuOH/CH3CN mixture (Case Study 4) further reduced the
MTSR.
formation of 25 (Case Study 7) did not meet the criteria.
Consequently, the reaction conditions investigated in Case
Study 4 were selected as the manufacturing conditions.
Based on the above modeling study, we subsequently
evaluated the process robustness with scale-down lab experi-
ments, by mimicking the commercial plant, to determine the
reaction parameters. The acceptable range of process
parameters such as flow rate, temperature, volume, and
operating time were appropriately defined by determining
the isolation yield and quality of 27 based on the criteria of
≥75% yield and ≥96% purity. Regarding the equipment factors
such as mixing, heat transfer, and fluctuation of flow rate which
would not correspond to the actual manufacturing system, we
designed a series of experiments under more severe conditions
to address these variables.
Furthermore, another improvement that was also applied
was the use of anhydrous dimethylamine for the aminolysis of
6 to avoid undesired competitive hydrolysis of the lactone ring
in 6. As a result, the overall yield of 27 from 6 was significantly
improved from 63% to 78%.
We summarized the changes to the manufacturing process to
prepare 3 in Scheme 9. Although the total number of reaction
steps was not decreased, the revised operations were
significantly simplified by telescoping the key reaction steps
such as the conversion of 27 to 3. Also, the overall yield for the
revised manufacturing process for the preparation of 3 from 6
was increased from 40% to 66%, which enabled an increase in
the amount of 3 generated by more than 1.5 times per batch.
In addition, the material cost of 3 was reduced by 40% due to
these improvements, which effectively reduced the cost to
produce edoxaban.
As a result, the yield of 30 was drastically increased and the
criteria for a suitable reaction were met even without the
installation of a heat exchanger. Although addition of a heat
exchanger could provide further improvement (Case Study 5),
the return on equipment investment would be expected to be
low. Therefore, we decided not to use a heat exchanger in this
process. The conditions that used a more dilute reaction (Case
Study 6) and an increased reaction temperature for the
H
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
method was dried under full vacuum, 165.5 g of dry rac-6 were
obtained in 51% yield. H NMR (CDCl3) 4.81 (t, J = 5.0 Hz,
1H), 4.41 (t, J = 4.4 Hz, 1H), 2.66 (d, J = 12.2 Hz, 1H), 2.08−
2.58 (m, 4H). 1.80−2.08 (m, 2H).
CONCLUSION
■
1
We have developed a highly efficient manufacturing route to
the key intermediate 3 in the synthesis of edoxaban. During the
investigation of an alternative route, we identified an enzymatic
resolution approach that avoided the use of a stoichiometric
amount of a chiral amine and repetitious diastereomeric salt
recrystallizations. Furthermore, in the course of the synthetic
route interrogation, an interesting cascade, cyclization reaction
was discovered and applied to design a more efficient route to
the cis-diamine intermediate. Although the method initially had
some challenges for the application to large scale processing,
the finalized procedure was reproducible and robust once the
application of a plug-flow reactor was incorporated into the
process. These technologies were highly effective in terms of
cost reduction and increase in productivity.
(1S,4S,5S)-4-Bromo-6-oxabicyclo[3.2.1]octan-7-one
(6). The previously prepared rac-6 wet cake (289.6 g) was
charged to water (480 mL) and stirred at 30 °C. Esterase
(Esterase AC “Amano”, 8.0 g) was added, and then the pH was
controlled between 7 to 8 using a 1 M potassium carbonate
aqueous solution (264 mL) with a pH-stat for 11 h. The
resulting solids were filtered, washed with water (200 mL), and
dried at room temperature to give crude 6 (86.4 g, 26.5% yield,
95% ee). The crude 6 (80 g) was charged to acetone (480 mL)
and dissolved at 50 °C. The resulting insoluble protein was
filtered, and the filtrate was concentrated to 320 mL. The
mixture was warmed to 50 °C, and water (480 mL) was added
to induce crystallization. The resulting slurry was cooled to 5
°C, stirred for 1 h, and filtered. The wet cake was washed with
40% acetone/water (200 mL) and dried to give 6 (75.0 g,
93.8% yield, 99% ee).
EXPERIMENTAL SECTION
■
General. All reactions were performed in an atmosphere of
nitrogen. Nuclear magnetic resonance (NMR) spectra were
obtained using a Bruker 400 MHz spectrometer in the
indicated solvents. Splitting patterns are indicated as follows: s,
singlet; d, doublet; t, triplet; m, multiplet; br, broad. Water
content was measured by Karl Fischer titration. All the yields
were calculated by weight without assay unless otherwise
noted. All reaction yields were quantified using a standard
sample to determine % weight by high performance liquid
chromatography (HPLC). Analytical conditions are described
below.
Compound 5 to 6. HPLC conditions (reaction monitoring
and evaluation of impurities). Buffer solution: Phosphate buffer
(pH = 2.5). Eluent A: Buffer and acetonitrile (7:3) v/v. Eluent
B: Buffer and acetonitrile (9:11) v/v. Column: X Bridge C8
(150 × 4.6) mm, 3.5 μm. Flow rate: 1.0 mL/min. Detector:
UV at 210 nm. Column oven temperature: 40 °C. Sample tray
temp: 5 °C. Injection volume: 10 μL. Run Time; 65 min.
Gradient Time (Min) Solution A (%)/Solution B (%): 0 min;
100:0, 10 min; 100:0, 35 min; 0:100, 45 min; 0:100, 60 min;
100:0.
Enantiomeric Purity of 6. HPLC conditions; Eluent:
premixed and degassed solution of n-hexane/2-propanol in the
ratio of 98:2 (%v/v). Column: CHIRALPAK AD-H (250 ×
4.6) mm, 5 μm. Flow rate: 1.0 mL/min. Detector: UV at 215
nm. Injection volume: 40 μL. Column oven temperature: 35
°C. Sample tray temperature: 25 °C. Run time: 30 min.
Compound 6 to 3. Column: L-Column C8 (4.6 mm I.D.
× 150 mm, 5 μm). Eluent: (0.02 M KH2PO4/CH3CN = 2:1)
+ 0.5 mol % Sodium dodecyl sulfate (SDS). Mode: Isocratic.
Column temp: 30 °C. Detector: UV at 210 nm. Flow rate: 1
mL/min. Time: 30 min. Sample cooler: 5 °C. Injection sample
was diluted with the same eluent.
4-Bromo-6-oxabicyclo[3.2.1]octan-7-one (rac-6).
Water (200 mL) and potassium hydroxide solution (48%
aqueous solution 123.9 mL, 1.590 mol) were treated by the
addition of rac-5 (200 g, 1.585 mol). 1,3-Dibromo-5,5-
dimethylhydantoin (226.7 g, 0.793 mol) was added over 3 h
while maintaining the reaction temperature below −10 °C and
then stirred at 0 °C for 1 h.
(1R,2R,4S)-2-[(tert-Buthoxycarbonyl)amino]-4-
[(dimethylamino)carbonyl]cyclohexylmethane-
sulfonate (11). Compound 6 (20 g, 0.098 mol) was mixed by
addition of acetonitrile (125 mL) and an aqueous solution of
dimethyl amine (50%, 35.2 g, 0.39 mol). The reaction mixture
was stirred at 10 °C for 15 h and then concentrated under
reduced pressure maintaining the temperature below 15 °C.
An aqueous ammonia solution (28%, 125 mL, 2.06 mol) was
added, and the mixture was stirred at 40 °C for 8 h and then
concentrated under reduced pressure. Deionized water (88
mL), di-tert-butyl dicarbonate (31.9 g, 0.146 mol), and NaOH
aqueous solution (48%, 20.3 g) were added and stirred for 2 h
at 40 °C. After reaction completion, 4-methyl-2-pentanone
(175 mL) was added, and the layers were separated. The
aqueous layer was extracted again with 4-methyl-2-pentanone
(175 mL), and the combined organic phases were concen-
trated under reduced pressure to a volume of 175 mL. 4-
Methyl-2-pentanone (75 mL) and methane sulfonyl chloride
(17.9 g, 0.156 mol) were added to the solution. Triethyl amine
(18.8 g, 0.186 mol) was added dropwise, and the reaction
mixture was stirred for 1 h at 25 °C. Methanol (43 mL) and
deionized water (63 mL) were added to the mixture, and
stirring was continued for 15 min. The organic layer was
washed with 5% sodium bicarbonate aqueous solution (50
mL), and the separated organic phase was concentrated under
reduced pressure to a volume of 100 mL. The resulting slurry
was stirred at 0 °C for 3 h and filtrated. The cake was washed
with 4-methyl-2-pentanone (25 mL) and dried under full
1
vacuum to give 11 (22.4 g, 62.9% yield). H NMR (CDCl3)
4.69−4.82 (m, 2H), 4.00−4.08 (m, 1H), 3.07 (s, 3H), 3.04, (s,
3H), 2.94 (s, 3H), 2.75−2.81 (m, 1H), 2.17−2.26 (m, 1H),
2.04−2.15 (m, 1H), 1.84−1.96 (m, 2H), 1.67−1.76 (m, 1H),
1.58−1.66 (m, 1H), 1.45 (s, 9H).
tert-Butyl{(1R,2S,5S)-2-amino-5-[(dimethylamino)-
carbonyl]cyclohexyl}carbamate Oxalate (3). Dodecylpyr-
idinium chloride (15.58 g, 0.055 mol), toluene (40 mL), and
NaN3 aqueous solution (13.56 g/40 mL, 0.207 mol) were
charged, and the mixture was stirred at 30 °C for 1 h. Toluene
(147 mL) was added, and a Dean−Stark trap was attached to
the flask. The mixture was refluxed under reduced pressure at
50 °C for 7 h to reduce the water content (0.05%,
specification: < 0.17%). The mesylate 11 (40 g, 0.110 mol)
was added at 20 °C, and the mixture was stirred at 70 °C for
Sodium sulfite (40.0 g 0.317 mol) was dissolved in water
(200 mL) and added to the reaction solution. This was stirred
at 30 °C for 1 h and filtered. The cake was washed with water
(400 mL) to give rac-6 (289.6 g as a wet cake) and then used
in the next step. When the wet cake prepared with the same
I
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
10 h. An aqueous solution of 2.5% sodium bicarbonate (120
mL) and toluene (28 mL) were added, the mixture was stirred
at 50 °C for 15 min, and the organic layer was separated. The
remaining aqueous layer was extracted with toluene (3 × 40
mL). The organic layers were combined and washed with
water (20 mL) and then concentrated to 120 mL under
reduced pressure. Toluene (40 mL), MeOH (4 mL), and Pd/
C (7.5%, 5.60 g) were added to the mixture, and an
ammonium formate (7.60 g, 0.121 mol) solution in toluene
(60 mL) and MeOH (44 mL) was added dropwise while the
reaction was stirred at 40 °C for 1 h and then filtrated. The
filtrate was concentrated under reduced pressure to 60 mL, and
acetonitrile (280 mL) and water (20 mL) were added. An
acetonitrile solution of oxalic acid (6.9%, 128.89 g, 0.099 mol)
was added dropwise to the solution at 50 °C and stirred for 3
h. The resulting slurry was filtrated at 30 °C. The wet
crystalline mass and CH3CN (280 mL) were charged to a flask,
and then the mixture was concentrated under reduced pressure
to remove water. The slurry was filtrated at 30 °C and dried
(CDCl3) 7.39 (s, 1H), 5.37 (d, J = 5.5 Hz, 1H), 4.72−4.75 (m,
1H), 3.99 (m, 1H), 3.10 (s, 3H), 3.06 (s, 3H), 2.94 (s, 3H),
2.82−2.86 (m, 1H), 2.35 (ddd, J = 4.8, 8.9, 14.4 Hz, 1H),
1.98−2.05 (m, 2H), 1.84−1.91 (m, 1H), 1.61−1.69 (m, 2H),
1.49 (s, 9H). HRMS: calculated for C15H29N3O8S2, M − H−
m/z 442.1318, found 442.1326.
tert-Butyl{(1R,2S,5S)-2-amino-5-[(dimethylamino)-
carbonyl]cyclohexyl}carbamate Oxalate (3) from 27.
Mesylate 27 (1430 kg, 3.224 kmol) was charged to a reactor,
followed by the addition of acetonitrile (6435 L) and
triethylamine (503 L, 3.579 kmol). The mixture was stirred
at 70 °C for 2 h. After the reaction completion, water (358 L)
and pyridine (1287 L, 15.98 kmol) were added, and the
mixture was stirred at 75 °C for 5 h. NaCl aqueous solution
(20%, 1430 L), water (200 L), toluene (14 300 L), and NaOH
aqueous solution (25%, 2145 L) were added, and the mixture
was stirred at 50 °C; then the layers were separated. The
organic layer was washed with NaCl aqueous solution (20%,
1430 L) and NaOH aqueous solution (25%, 286 L). The
organic layer was concentrated under reduced pressure to a
volume of 4290 L, and then acetonitrile (8580 L) was added
and filtered to give the solution of 3. Oxalic acid (406 kg, 4.510
kmol), water (601 L), and acetonitrile (10 100 L) were
charged to another reactor, followed by the addition of 3
solution at 35 °C for 2 h. The mixture was stirred at 50 °C for
2 h, then cooled to 25 °C, and filtrated. The cake was washed
with an aqueous acetonitrile solution (93%, 5434 L) to give
the crude crystalline 3 as a hydrate. Then the crude 3 was
charged into acetonitrile (10 010 L), and the mixture was
concentrated under reduced pressure to remove most of the
water. The slurry was filtrated at 30 °C and then dried under
full vacuum at 60 °C to give 3 (1010.8 kg, 2.692 kmol, 83.5%
yield) as a white crystalline solid.
1
under full vacuum to give 3 (26.33 g, 0.0701 mol, 63.8%). H
NMR (CDCl3) 4.10 (br, 1H), 3.32 (d, J = 12.2 Hz, 1H), 2.96
(m, 3H), 2.80 (t, J = 12.4 Hz, 1H), 2.77 (s, 3H), 1.72−1.83
(m, 3H), 1.63 (t, J = 22.7 Hz, 1H), 1.37−1.49 (m, 2H), 1.30
(s, 9H).
Preparation of Burgess Reagent 25. Chlorosulfonyl
isocyanate (737 kg, 5.207 kmol) and t-BuOH (465 kg, 6.274
kmol) in acetonitrile (3897 L) were reacted through the plug-
flow system, and the mixture was continuously added into a
mixture of triethylamine (2117 L, 15.06 kmol) and acetonitrile
(1477 L) while maintaining the temperature below 5 °C for 13
h.
(1R,2R,4S)-2-{[(tert-Butoxycarbonyl)sulfamoyl]-
amino}-4-(dimethylcarbamoyl)cyclohexyl Methane-
sulfonate (27) from 6.17 (1S,4S,5S)-4-Bromo-6-oxabicyclo-
[3.2.1]octan-7-one (890 kg, 4.340 kmol) was charged followed
by dimethylamine in THF (24%, 2092 L). The mixture was
stirred at 25 °C for 14 h. NaCl (409 kg), citric acid (584 kg),
water (2252 L), and ethyl acetate (3560 L) were added, and
the layers were separated. The aqueous layer was extracted
with ethyl acetate (2 × 1780 L). The organic phases were
combined and concentrated under reduced pressure to a
volume of 2225 L. Water (1780 L) was added, and the solution
was concentrated under reduced pressure to a volume of 2225
L. Water (1638 L), ammonia aqueous solution (28%, 5563 L,
82.37 kmol), and NaOH aqueous solution (25%, 547 L) were
added, and the mixture was stirred at 40 °C for 4 h and
concentrated under reduced pressure to a volume of 4450 L to
afford a 9 solution. The solution was added to the previously
prepared 25 solution, and the mixture was stirred at 5 °C for 1
h. Citric acid (334 kg) and ethyl acetate (9790 L) were added,
and the layers were separated. The aqueous layer was extracted
with ethyl acetate (3193 L). The organic layers were combined
and washed with a NaCl aqueous solution (20%, 3382 L), and
the solvent was replaced to give an acetonitrile solution of 26
(2670 L). Methane sulfonyl chloride (488 kg, 4.261 kmol) and
4-methylmorpholine (531 kg, 5.250 kmol) were added, and the
mixture was stirred at 5 °C for 3 h. Sulfuric acid (5% aqueous
solution, 757 L) was added, and the mixture was transferred
into water (12 990 L). The slurry was stirred at 5 °C for 3 h
and then filtrated. The cake was washed with aqueous
acetonitrile (28%, 2848 L) and then dried under full vacuum
at 50 °C to give the corresponding mesylate 27 (1529.9 kg,
3.449 kmol, 79.5% yield) as a white crystalline solid. 1H NMR
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.oprd.8b00413.
■
S
Optimization studies of enzymatic reaction, confirma-
tory experiments, effect of hydrodynamics and process
modeling in plug-flow reactor (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: michida.makoto.wp@daiichisankyo.co.jp.
■
ORCID
Makoto Michida: 0000-0003-1281-3300
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Mr. Tomohiro Shiohara, Yasuhisa
Kuwahara, Takuma Nishizawa, and the staff of Manufactur-
ing/Engineering Department of the Odawara plant (Daiichi
Sankyo Chemical Pharma Co., Ltd.) for their valuable support
during our scale up runs. In addition, we thank Dr. David
Conlon (Daiichi Sankyo, Inc.) for helpful discussions during
the preparation of the manuscript.
REFERENCES
■
(1) (a) Levy, J. H.; Spyropoulos, A. C.; Samama, C. M.; Douketis, J.
Direct Oral Anticoagulants: New Drugs and New Concepts. JACC:
J
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Cardiovasc. Interventions 2014, 7, 1333. (b) Pinto, D. J. P.; Orwat, M.
J.; Koch, S.; Rossi, K. A.; Alexander, R. S.; Smallwood, A.; Wong, P.
C.; Rendina, A. R.; Luettgen, J. M.; Knabb, R. M.; He, K.; Xin, B.;
Wexler, R. R.; Lam, P. Y. S. Discovery of 1-(4-Methoxyphenyl)-7-oxo-
6-(4-(2-oxopiperidin-1-yl)phenyl)-4,5,6,7-tetrahydro-1H-pyrazolo-
[3,4-c]pyridine-3-carboxamide (Apixaban, BMS-562247), a Highly
Potent, Selective, Efficacious, and Orally Bioavailable Inhibitor of
Blood Coagulation Factor Xa. J. Med. Chem. 2007, 50, 5339.
(c) Roehrig, S.; Straub, A.; Pohlmann, J.; Lampe, T.; Pernerstorfer, J.;
Schlemmer, K.-H.; Reinemer, P.; Perzborn, E. Discovery of the Novel
Antithrombotic Agent 5-Chloro-N-({(5S)-2-oxo-3- [4-(3-oxomor-
pholin-4-yl)phenyl]-1,3-oxazolidin-5-yl}methyl)thiophene-2-carboxa-
mide (BAY 59-7939): An Oral, Direct Factor Xa Inhibitor. J. Med.
Chem. 2005, 48, 5900. (d) Norbert, H.; Hauel, N. H.; Nar, H.;
Priepke, H.; Ries, U.; Stassen, J.-M.; Wienen, W. Structure-Based
Design of Novel Potent Nonpeptide Thrombin Inhibitors. J. Med.
Chem. 2002, 45, 1757.
Synthesis of Pharmaceutical Products. Org. Process Res. Dev. 2016, 20,
2. (c) Gutmann, B.; Cantillo, D.; Kappe, C. O. Continuous Flow
Technology A Tool for the Safe Manufacturing of Active
Pharmaceutical Ingredients. Angew. Chem., Int. Ed. 2015, 54, 6688.
(d) Teoh, S. K.; Rathi, C.; Sharratt, P. Practical Assessment
Methodology for Converting Fine Chemicals Processes from Batch
to Continuous. Org. Process Res. Dev. 2016, 20, 414.
(15) (a) Bourne, J. R. Mixing and the Selectivity of Chemical
Reactions. Org. Process Res. Dev. 2003, 7, 471. (b) Aubin, J.; Ferrando,
M.; Jiricny, V. Current methods for characterising mixing and flow in
microchannels. Chem. Eng. Sci. 2010, 65, 2065. (c) Commenge, J. M.;
Falk, L. Villermaux−Dushman protocol for experimental character-
ization of micromixers. Chem. Eng. Process. 2011, 50, 979.
(16) Nagaki, A.; Togai, M.; Suga, S.; Aoki, N.; Mae, K.; Yoshida, J.
Control of Extremely Fast Competitive Consecutive Reactions using
Micromixing. Selective Friedel−Crafts Aminoalkylation. J. Am. Chem.
Soc. 2005, 127, 11666.
(17) Nakamura, Y.; Michida, M.; Kaneda, T. Preparation method of
optically active diamine compound, Patent US20160016974, Jul. 21,
2015.
(2) Hirsh, J.; Fuster, V. Guide to anticoagulant therapy. Part 2: Oral
anticoagulants. Circulation 1994, 89, 1469.
(3) (a) Giugliano, R. P.; Ruff, C. T.; Braunwald, E.; Murphy, S. A.;
Wiviott, S. D.; Halperin, J. L.; Waldo, A. L.; Ezekowitz, M. D.; Weitz,
̌
J. I.; Spinar, J.; Ruzyllo, W.; Ruda, M.; Koretsune, Y.; Betcher, J.; Shi,
M.; Grip, L. T.; Patel, S. P.; Patel, I.; Hanyok, J. J.; Mercuri, M.;
Antman, E. M.; Investigators, E. A.-T. Edoxaban versus warfarin in
patients with atrial fibrillation. N. Engl. J. Med. 2013, 369, 2093.
(b) The Hokusai-VTE Investigators. Edoxaban versus Warfarin for
the Treatment of Symptomatic Venous Thromboembolism. N. Engl. J.
Med. 2013, 369, 1406.
(4) Ohta, T.; Komoriya, S.; Yoshino, T.; Uoto, K.; Nakamoto, Y.;
Naito, H.; Mochizuki, A.; Nagata, T.; Kanno, H.; Haginoya, N.;
Yoshikawa, K.; Nagamochi, M.; Kobayashi, S.; Ono, M. Diamine
derivatives. Patent WO 2003/000680, Jan. 3, 2003.
(5) Sato, K.; Kawanami, K.; Yagi, T. Optically active diamine
derivative and process for producing the same. Patent
US20090105491, Apr. 23, 2009.
(6) Schwartz, H. M.; Wu, W.-S.; Marr, P. W.; Jones, J. B. Predicting
the enantiomeric selectivity of chymotrypsin. Homologous series of
ester substances. J. Am. Chem. Soc. 1978, 100, 5199.
(7) Enzyme screening was conducted using Chiral Enzyme
Spectrum “Amano” which was provided from Amano Enzyme Inc.
(8) Furman, D.; Dubnikova, F.; van Duin, A. C. T.; Zeiri, Y.; Kosloff,
R. Reactive Force Field for Liquid Hydrazoic Acid with Applications
to Detonation Chemistry. J. Phys. Chem. C 2016, 120, 4744.
(9) (a) Langlois, N.; Moro, A. Regio- and Stereoselective Opening
of Oxiranes through Neighbouring Group Participation: Stereo-
controlled Synthesis of Enantiopure Hydroxylated Oxazolidin-2-ones.
Eur. J. Org. Chem. 1999, 1999, 3483. (b) Crich, D.; Hu, T.; Cai, F.
Does Neighboring Group Participation by Non-Vicinal Esters Play a
Role in Glycosylation Reactions? Effective Probes for the Detection of
Bridging Intermediates. J. Org. Chem. 2008, 73, 8942. (c) Agami, C.;
Couty, F. The reactivity of the N-Boc protecting group: an underrated
feature. Tetrahedron 2002, 58, 2701.
1
(10) The structure of 20 was assigned by H−1H COSY spectra.
1
(11) NMR spectra of aziridine 23: H NMR (CDCl3) 7.35−7.39
(m, 5H), 5.22 (s, 2H), 4.16 (q, J = 7.0 Hz, 2H), 3.21−3.23 (m, 1H),
3.10−3.13 (m, 1H), 2.41−2.45 (m, 1H), 1.91−1.98 (m, 2H), 1.80−
1.86 (m, 1H), 1.72−1.75 (m, 1H), 1.34−1.45 (m, 1H), 1.26 (t, J =
7.0 Hz, 3H).
(12) (a) Kawato, H.; Miyazaki, M.; Sugimoto, Y.; Naito, H.;
Okayama, T.; Soga, T.; Uoto, K. Imidazothiazole derivatives, Patent
WO2008072655. (b) Kurokawa, T.; Kim, M.; Du Bois, J. Synthesis of
1,3-diamines through rhodium-catalyzed C-H insertion. Angew.
Chem., Int. Ed. 2009, 48, 2777. (c) Olson, D. E.; Su, J. Y.; Roberts,
D. A.; Du Bois, J. J. Am. Chem. Soc. 2014, 136, 13506.
(13) Crimmins, M. T.; Kim-Meade, A. S. Isobutene, e-EROS
Encyclopedia of Reagents for Organic Synthesis, 2001, 1.
(14) (a) Anderson, N. G. Using Continuous Processes to Increase
Production. Org. Process Res. Dev. 2012, 16, 852. (b) Porta, R.;
Benaglia, M.; Puglisi, A. Flow Chemistry: Recent Developments in the
K
DOI: 10.1021/acs.oprd.8b00413
Org. Process Res. Dev. XXXX, XXX, XXX−XXX