A one-pot asymmetric synthesis of a N-acylated 4,5-dihydropyrazole, a key intermediate of thrombin inhibitor AZD8165

Article abstract of DOI:10.1021/op500134e

A short, chromatography-free, and scalable synthetic route to thrombin inhibitor 1, the active metabolite of the propionic ester prodrug AZD8165, has been developed. The key synthetic step involved cycloaddition of TMS-diazomethane and ethyl acrylate to give an intermediate racemic dihydropyrazole which was reacted with enantiomerically pure 4-fluoro mandelic acid chloride in a one-pot dynamic kinetic resolution (DKR) process.

Full text of DOI:10.1021/op500134e

Article
pubs.acs.org/OPRD
A One-Pot Asymmetric Synthesis of a N‑Acylated 4,5-
Dihydropyrazole, A Key Intermediate of Thrombin Inhibitor
AZD8165
̈
Staffan Karlsson,* Jonas Branalt, Maria Olwegard Halvarsson, and Joakim Bergman
̊
̊
Medicinal Chemistry, Cardiovascular & Metabolic Diseases iMed, AstraZeneca R&D Molndal, S-431 83 Molndal, Sweden
̈
̈
S
* Supporting Information
ABSTRACT: A short, chromatography-free, and scalable synthetic route to thrombin inhibitor 1, the active metabolite of the
propionic ester prodrug AZD8165, has been developed. The key synthetic step involved cycloaddition of TMS−diazomethane
and ethyl acrylate to give an intermediate racemic dihydropyrazole which was reacted with enantiomerically pure 4-fluoro
mandelic acid chloride in a one-pot dynamic kinetic resolution (DKR) process.
utilizing a chiral auxiliary2 (medicinal chemistry route 1,
INTRODUCTION
■
Scheme 1), or a chiral pool approach (medicinal chemistry
route 2) was chosen.3 The racemic mandelic acid building
block A was used in the coupling with BC which resulted in a
diastereomeric mixture of 1. This was subsequently separated
by chromatography to give the stereoisomerically pure
compound 1. Using these approaches, the medicinal chemistry
group supplied the project with the API 1 for the initial in vitro
screening activities.
For large-scale purposes these medicinal chemistry ap-
proaches were deemed unsuitable mainly due to the high
energetic properties of tetrazole C (R2 = H).4 Moreover, the
synthetic route to C included several challenging chemical
transformations and chromatographic purifications.5 Conse-
quently, a different strategy of building block assemblage, in
which C was introduced as late as possible in the synthetic
sequence, was desired.
Thrombin plays a central role in the generation of a thrombus
by converting soluble fibrinogen to insoluble fibrin which
stabilizes the clot during the wound healing process.1 The
inhibition of thrombin will break down the clot and prevent
undesired vascular events. In the search for novel direct
thrombin inhibitors with improved properties compared to the
historically used anticoagulant Warfarin, we identified dihy-
dropyrazole derivative 1, the active metabolite of the
propionate ester prodrug AZD8165 (Figure 1).1b In order to
support preclinical and clinical studies of AZD8165, several
hundred grams of 1 were needed.
RESULTS AND DISCUSSION
■
Our primary goal for the first scale-up campaign was to quickly
develop a short, scalable, chromatography-free, and safe
process. We found that the BOC protected amine C (R2 =
BOC) was safer to handle compared with the primary amine C
(R2 = H).6,7 The plan was to perform an in situ deprotection of
C (R2 = BOC) to give the free amine and then use this as a
solution in the subsequent step without isolation.
Figure 1. Retrosynthetic analysis and synthetic strategies for the
thrombin inhibitor 1.
Although an asymmetric approach to the dihydropyrazole
building block B was desired, we wanted to avoid the use of
chiral auxiliaries. However, to the best of our knowledge, with
the exception of a few special cases,8 enantioselective syntheses
of dihydropyrazoles of type B mediated by chiral catalysts were
not known in the literature. Instead, we aimed for obtaining the
racemate of the dihydropyrazole B and then find conditions to
resolve the enantiomers through salt formation. Alternatively,
we envisioned to separate the diastereomers formed from the
reaction of the enantiomerically pure mandelic acid building
Retrosynthetically, compound 1 is conveniently disconnected
via strategic amide bond linkages, revealing building blocks A,
B, and C (R2 = H), or derivatives thereof, as suitable coupling
partners. The medicinal chemistry program towards this series
of compounds focused on variation of building block A. In the
synthetic strategy, A was introduced as the last step by means of
an acylation reaction with BC, which in turn was prepared from
dihydropyrazole B and tetrazole C (R2 = H) via an amide
coupling reaction. The detailed medicinal chemistry routes to
the API 1 are outlined in Scheme 1. To obtain the pure
enantiomer of the dihydropyrazole building block B, a strategy
Received: April 22, 2014
Published: August 5, 2014
© 2014 American Chemical Society
969
dx.doi.org/10.1021/op500134e | Org. Process Res. Dev. 2014, 18, 969−975

Organic Process Research & Development
Article
Scheme 1. Medicinal chemistry approaches to API 1
block A and the racemic dihydropyrazole B at a later stage. For
this latter strategy, enantiomerically pure mandelic acid A was
needed. Although various large-scale enzymatic methods have
been reported for the preparation of enantiomerically pure
mandelic acid derivatives,9 racemic 4-fluoro mandelic acid is a
readily available commercial compound, and our strategy was to
find efficient ways to resolve the racemate.
Scheme 3. One-pot synthesis of the diastereomeric
dihydropyrazoles 6 and 7
Large-Scale Synthesis. A literature survey showed that
racemic 4-fluoromandelic acid could be resolved into pure
enantiomers through salt crystallization with enantiomerically
pure α-phenylethylamine.10 After further optimization, a
practical large-scale method for this resolution was established,
furnishing >400 g (25% yield) of (R)-2 with ee >99% (Scheme
2). Synthesis of TMS-protected acid chloride (R)-3 was
Scheme 2. Resolution of 4-fluoromandelic acid and
transformation to acid chloride (R)-3
4,14 we found it impractical to perform a solvent swap
operation on a large scale. Instead, our plan was to use CH2Cl2
also for the cycloaddition reaction.
To our delight, the cycloaddition between ethyl acrylate and
TMS-diazomethane proceeded smoothly at 20 °C in CH2Cl2
with consistent results. Desilylation of the intermediate 4 was
accomplished using an excess of TFA at room temperature
which gave a clean and full conversion to 5. However, isolation
of the highly water-soluble free base of 5 was difficult. Previous
reports on its isolation and purification involved silica gel
chromatography, which we considered not to be practical on a
large scale.13b Dihydropyrazoles of type 5 are also known to
slowly oxidize to the corresponding pyrazole, but this can be
prevented by N-acylation.13a We reasoned that, in order to
develop a scale-up friendly method and to avoid the isolation/
oxidation issues of 5, this should preferably be telescoped
directly into the acylation step. Therefore, an attempt was made
in which 5 was directly acylated with acid chloride (R)-3
(Scheme 3). After treatment of the crude mixture of 5 with
pyridine and acid chloride (R)-3, a diastereomeric mixture of 6
and 7 was obtained in moderate yield. With the aim to improve
the yield, we changed the order of the addition of base (i.e.,
adding pyridine last). Interestingly, it was found that the
presence of an amine base was not necessary for this acylation
achieved by a one pot TMS-protection/acid chloride
generation procedure,11 and then used directly in the
subsequent step as a solution in CH2Cl2.
Several methods are available for the synthesis of racemic
dihydropyrazoles of type B (Figure 1).12 We found the 1,3-
dipolar cycloaddition between commercially available TMS-
diazomethane and acrylates in nonpolar solvents such as
heptanes and toluene, to be the most attractive approach
(Scheme 3).13 The reaction is known to proceed via
cycloadduct 4, and after a solvent swap to CH2Cl2 followed
by treatment with a strong acid such as TFA, dihydropyrazole 5
is obtained.13b Because of the unstable nature of intermediate
970
dx.doi.org/10.1021/op500134e | Org. Process Res. Dev. 2014, 18, 969−975

Organic Process Research & Development
Article
reaction. In fact, a much cleaner reaction profile was observed
by excluding the base. Acylation of the dihydropyrazole under
the acidic conditions used might be facilitated by the
neighbouring effect of the α-nitrogen atom which could assist
in accepting a proton, thus making the other nitrogen available
for reaction with acid chloride (R)-3 (Scheme 4).15,16
and acid chloride (R)-3. We anticipated that using a larger
excess of TFA would increase the rate of the dynamic
racemization of the intermediate dihydropyrazole 5, resulting
in a higher yield of 6. However, although a slightly better yield
of 6 was obtained using 2 equiv of TFA, this was probably due
to a cleaner reaction profile (entry 5 vs 3). To have a successful
DKR, the product distribution of 6/7 must not change during
the course of the reaction. After reaction completion we found
that the initial ratio was preserved also after 20 h of stirring.
This was also the case when the initial crude reaction mixture of
6/7 was doped with 7 to give a 40/60 diastereomeric ratio and
then allowed to stir for 20 h. Also, we found that the product
distribution 6/7 was the same after 20%, 50%, and 100% of the
acid chloride (R)-3 had been added. Thus, these observations
also support the DKR of 5 proposed above while leaving 6/7
untouched.17 Diastereomers 6/7 were separated using silica gel
chromatography. Later, we found that compound 6 could be
obtained as a pure diastereomer after a single crystallization of
the crude product mixture from EtOAc/heptane.
Scheme 4. Proposed mechanism for the DKR of the
dihydropyrazole under strongly acidic conditions
Having established a one-pot method for the synthesis of
dihydropyrazole 6, we were ready to scale up the reaction.
However, due to the high toxicity of TMS−diazomethane,18
a
method for the detection of this reagent in the crude mixtures
had to be established to ensure full conversion prior to work-
up. This would enable us to handle this reagent safely also on a
large scale. We found that the amount of TMS-diazomethane in
the reaction mixture could be monitored both by in situ FT-IR
(2066 cm−1) and by 1H NMR (2.6 ppm, s, CDCl3)
measurement (Figure 2). Using a slight excess of the ethyl
acrylate, no residual unreacted TMS-diazomethane was
detected after 20 h of stirring at 20 °C.
Using the one-pot approach as described above (Scheme 3),
345 g (51% yield based on TMS-diazomethane) of
dihydropyrazole 6 with ee and dr >99% was isolated after
one crystallization from EtOAc/heptane.
Upon acylation of the dihydropyrazole 5 (200 mol %) with
acid chloride (R)-3 (100 mol %), a 70:30 diastereomeric
mixture of 6 and 7, in favor of the desired isomer 6, was
obtained, indicating that a kinetic resolution of the dihydropyr-
azole had occurred (Scheme 4). Surprisingly, the diastereose-
lectivity was the same using a 1:1 ratio of 5 and (R)-3. Thus,
during the strongly acidic conditions used in the acylation step,
dihydropyrazole 5 undergoes a dynamic racemization.
Combined with the kinetic resolution of 5, upon reaction
with acid chloride (R)-3, a good overall assay yield of the
desired product (6) was obtained. We decided to study this
reaction in more detail to find better conditions for this
transformation (Table 1).
To obtain the desired API 1, hydrolysis of ester 6 followed by
coupling with the tetrazole building block 10 was required.
Unfortunately, in the hydrolysis of 6 using standard alkaline
reagents such as NaOH or LiOH, as much as 30%
epimerization of the α-CH position with respect to the
carboxylic acid was detected (Table 2).19 Moreover, hydrolysis
of the amide bond of 8 to give (R)-2 was a major competing
side reaction. Attempts to perform a direct aminolysis of the
ethyl ester 6 with amine 10 using NaCN as a catalyst20 gave
only trace amounts of 1 even at elevated temperatures.
However, due to the explosive properties of 10, heating of
the reaction mixture should be avoided. It was clear that none
of the above methods were suitable for further scale up.
To our delight, when the mild hydrolysis method mediated
by Et3N/LiBr, previously developed by us, was applied,21 the
hydrolysis proceeded smoothly within a few hours at room
temperature to give carboxylic acid 8 in almost quantitative
yield with only 3% of Epi-8 detected (Table 2, entry 12). This
hydrolysis method gave consistent results also on a larger scale,
and no amide hydrolysis of 8 was observed.
In the subsequent coupling step, amine 10 was used as a
solution in EtOAc after deprotection of precursor 9 using HCl
(Scheme 5).22 By using TBTU as coupling reagent we could
use the wet crude EtOAc extract of 10 in the coupling with 8. It
was found that the product 1 had a very low solubility in EtOAc
and precipitated out from the reaction mixture. Thus, after
filtration of the product suspension, both the potential
explosive HOBt23 formed and the high energetic unreacted
It was found that the temperature did not have a significant
effect on the diastereomeric ratio 6:7 (entries 1−5). Moreover,
the loading of (R)-3 had no influence on the diastereomeric
ratio (compare entries 5 and 6), which is in line with the DKR
proposed. To generate less waste, the reaction should thus be
performed using an equimolar amount of dihydropyrazole 5
Table 1. Acylation of the crude dihydropyrazole 5 with acid
a
chloride (R)-3
b
c
entry temp. (°C) TFA (equiv) (R)-3 (equiv) dr (6:7)
yield 6
1
2
3
4
5
6
−78
−40
20
1.1
1.1
1.1
2.0
2.0
2.0
0.5
0.5
0.5
0.5
0.5
1.0
70:30
73:27
71:29
76:24
74:26
70:30
55%
57%
55%
64%
68%
58%
0
20
20
a
To a solution of ethyl acrylate (2.6 g, 26.3 mmol) in CH2Cl2 (40 mL)
was added a solution of TMS−diazomethane (2 M in hexanes, 13.1
mL, 26.2 mmol). The mixture was stirred at room temperature until
full conversion to cycloadduct 4 was obtained. This was followed by
the slow addition of TFA. After 30 min of stirring, acid chloride (R)-3
(50 or 100 mol %) was slowly added during 15 min. When reaction
was complete, the reaction was worked up with MeOH/water then
NaHCO3 (aq.) followed by concentration to give the crude
b
1
diastereomeric mixture of 6/7. Determined by H NMR through
c
1
integration. Assay yield (crude mixture) by H NMR based on the
limiting reagent and using benzyl benzoate as an internal standard.
971
dx.doi.org/10.1021/op500134e | Org. Process Res. Dev. 2014, 18, 969−975

Organic Process Research & Development
Article
Figure 2. FT-IR spectra of the reaction between TMS-diazomethane and ethyl acrylate.
Table 2. Hydrolysis of ester 6 using various methods
a
b
entry
method
solvent
base
dr (8:Epi-8)
amide hydrolysis (%)
c
1
A
A
A
A
A
A
A
A
A
A
A
B
MeOH
MeOH
MeOH
MeOH
THF
LiOH
NaOH
KOH
(78:22) 73:27
2
c
2
(81:19) 71:29
0
3
72:28
68:32
100:0
92:8
0
4
CsOH
LiOH
NaOH
CsOH
LiOH
NaOH
KOH
0
5
49
4
6
THF
7
THF
89:11
31
29
8
c
8
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
(93:7) 100:0
9
88:12
84:16
80:20
10
11
12
0
CsOH
Et3N
15
0
c
(97:3) 97:3
a
Method A: To a 0.2 M solution of 6 (20−60 mg) in the specified solvent was added the specified base (200 mol %) as a 0.4 M solution in water.
The mixture was stirred for 22 h, quenched by the addition of 1 M HCl and extracted with MTBE. The organic layer was then concentrated to give
the crude mixture of carboxylic acid 8. Method B: To a mixture of LiBr (111 mg, 1.27 mmol) and Et3N (89 μL, 0.64 mmol) in water (5 mg) and
CH3CN (0.75 mL) was added 6 (75 mg, 0.25 mmol). The suspension was stirred for 22 h, quenched by the addition of 1 M HCl, and extracted with
b
MTBE. The organic layer was then concentrated to give the crude mixture of carboxylic acid 8. Determined by chiral HPLC and 1H NMR analyses.
c
Diastereomeric ratio (dr) after 1 h of stirring.
activation with TBTU, nor by applying forcing conditions on
ester 6 (e.g., pTSA, toluene, reflux).
Scheme 5. Mild hydrolysis of 6 and coupling of 8 with
tetrazole building block 10
Figure 3. Tentative formation of the lactone as a side reaction.
Although it is known that, upon activation of chiral
carboxylic acids of type 8, oxazolone formation can occur
which often results in epimerization of the α-CH position with
respect to the carboxylic acid;27 this was not observed in our
case.
amine 10 could be safely washed out. The filter cake was dried
to give 182 g of 1 in 59% yield.24,25
In contrast to the observations by Nelson et al. using a
proline derivative,26 in the coupling step of 8 with 10, we did
not observe any lactone formation of 8 (Figure 3), neither by
CONCLUSION
■
A new chromatography-free route to the thrombin inhibitor 1
was discovered and further developed to a practical method for
972
dx.doi.org/10.1021/op500134e | Org. Process Res. Dev. 2014, 18, 969−975

Organic Process Research & Development
Article
large-scale application. Compared with the long (8−12 steps)
linear medicinal chemistry routes to the API 1 in 5−10%
overall yield, a considerably shorter convergent synthesis with
better overall yield [3 steps, 29% from (R)-2] was developed.
The key step was the discovery of a one-pot DKR of an
intermediate racemic dihydropyrazole (5) mediated by
acylation with enantiomerically pure acid chloride (R)-3
under strongly acidic conditions. The safety concern using
the potentially explosive building block 10 was solved through a
telescoping sequence from a safer precursor (9). Problematic
epimerization during hydrolysis of ester 6 was minimized using
the mild LiBr/Et3N method developed by us previously.21 The
developed synthesis was demonstrated on a multigram scale
delivering API 1 in >99% ee and >99% dr.
fluorophenyl)-2-hydroxyacetic acid (R)-2 (392 g, 2.3 mol) and
CH2Cl2 (4 L) followed by the addition of pyridine (383 g, 4.84
mol) and 4-dimethylaminopyridine DMAP (14.1 g, 0.12 mol).
The jacket temperature was set at 20 °C, and to the clear
homogeneous solution was added trimethylsilyl chloride TMS-
Cl (526 g, 4.84 mol) during 30 min. The reaction temperature
increased to 30 °C during this addition. The suspension formed
was stirred for 20 h at 20 °C followed by cooling to 0 °C. DMF
(17 g, 0.23 mol) was added followed by slow addition of oxalyl
chloride (301 g, 2.4 mol) dissolved in CH2Cl2 (60 mL) during
25 min. The reaction temperature increased to +7 °C during
this addition. After being stirred for an additional 1 h at 20 °C,
the titled crude acid chloride was used as such in the next step.
The yield was assumed to be quantitative.
(S)-Ethyl-1-((R)-2-(4-fluorophenyl)-2-hydroxyacetyl)-
4,5-dihydro-1H-pyrazole-5-carboxylate (6). With a jacket
temperature set at 20 °C, a 25 L reactor was charged with
dichloromethane (4 L) and ethyl acrylate (288 g, 2.88 mol).
Trimethylsilyldiazomethane (2 M in hexanes, 1150 mL, 2.30
mol) was slowly added during 30 min. The reaction
temperature increased to 22 °C during this addition. The
EXPERIMENTAL SECTION
■
General. All materials were purchased from commercial
suppliers and used as such without further purification. The
reactions were performed under an atmosphere of nitrogen in
reactors equipped with an overhead stirrer and connected to an
external heating Huber aggregate. In process controls (IPCs)
were recorded using either HPLC or 1H NMR analyses on the
crude mixtures. Diastereomeric ratios were determined using
1H NMR integration. Assays of final products and crude
mixtures were determined by 1H NMR integration using benzyl
benzoate as internal standard. High-resolution mass spectrom-
etry was performed on a Waters XEVO qTOF instrument, mass
precision 5 ppm. NMR measurement was performed on a
Bruker Avance III spectrometer. LC/MS analyses were
recorded on a Waters ZMD instrument, LC column Terra
MS C8 (Waters); detection with a HP 1100 MS-detector diode
array. In situ FT-IR analysis was recorded using a React IR 15
instrument coupled to a DS 9.5 mm AgX SiComp probe.
(R)-2-(4-Fluorophenyl)-2-hydroxyacetic acid [(R)-2].
Racemic 2-(4-fluorophenyl)-2-hydroxyacetic acid (1.75 kg,
10.3 mol) and EtOH (99.9%, 10 L) was charged to a 25 L
reactor. The mixture was heated to reflux. To the homogeneous
solution was added (R)-(+)-1-phenylethylamine (0.89 kg, 7.3
mol). The resulting clear pale yellow solution was allowed to
attain 60 °C with 1 °C/min. Within 1 h of stirring,
crystallization had initiated, and the mantle temperature was
set at 20 °C with 0.5 °C/min. The suspension was then stirred
for 16 h followed by filtration and washing with EtOH (2 L).
The wet filter cake was resuspended in EtOH (99.9%, 9 L) and
then heated to reflux. The homogeneous solution was then
allowed to attain 69 °C. Seeding crystals were added which
initiated a slow crystallization. After being stirred at 69 °C for
30 min the temperature was set at 20 °C with 0.3 °C/min, and
the mixture was stirred for an additional 22 h. The mixture was
filtered to give the wet salt (1.1 kg) with ee 95% (chiral HPLC).
One more crystallization from EtOH (99.9%, 8 L) using the
conditions above gave the stereoisomerically pure salt (ee
>99% free acid, 794 g, 26% yield). This salt was partitioned
between MTBE (2.5 L) and 3.8 M HCl (aq., 1 L). The aqueous
layer was extracted with MTBE (2 × 1 L), and the pooled
organic layer was washed with water (1.5 L) followed by
concentration and drying of the solid (399 g, 2.34 mol, 23%
yield) under reduced pressure. An additional 37 g (0.22 mol,
2% yield) of the titled acid was obtained through extraction of
1
homogeneous mixture was stirred at 20 °C for 19 h. H NMR
and FT-IR analysis of the crude solution showed full conversion
of the TMS−diazomethane to the intermediate unstable
cycloadduct 4. The jacket temperature was set at −30 °C.
Trifluoroacetic acid (656 g, 5.7 mol) was slowly added during
35 min. The reaction temperature increased to −13 °C during
this addition. The jacket temperature was set at 20 °C, and a
crude suspension of (R)-2-(4-fluorophenyl)-2-
(trimethylsilyloxy)acetyl chloride (R)-3 (600 g, 2.3 mol) in
dichloromethane (4 L) was added during 110 min; the reaction
temperature was kept below 20 °C during this addition. After
an additional 40 min of stirring, the reaction was quenched by
the addition of EtOH (500 mL) and water (500 mL). The
mixture was stirred for an additional 40 min followed by the
addition of additional water (2.5 L). The aqueous red layer (4
L) was separated from the organic one (orange 10.5 L). The
organic layer was extracted with an aqueous solution of
NaHCO3 (8%, 2.5 L) and then concentrated to a yellow solid.
1H NMR analysis of the crude mixture showed showed a
diastereomeric ratio (6:7) of 68:32. The crude mixture was
suspended in EtOAc (2 L) and heptane (2 L). The mixture was
heated to reflux. The resulting homogeneous solution was
cooled to 65 °C, and seeding crystals were added which
resulted in an immediate crystallization. The jacket temperature
was set at 20 °C with 1 °C/min, and the suspension was then
stirred for an additional 24 h followed by filtration and washing
of the solid with 50% EtOAc/heptane (3 × 300 mL). The solid
was then dried under reduced pressure furnishing 345 g (1.17
mol, 51% yield) of the titled colorless compound, assay 100%
1
w/w by H NMR, dr >99:1, ee >99%. HRMS (ESI+) m/e
1
295.1097 [(M + H)+, calcd for C14H16FN2O4 295.1089]. H
NMR (400 MHz, CDCl3) δ 1.28 (t, J = 7.1 Hz, 3H); 2.92 (ddd,
J = 18.8, 5.6, 1.8 Hz, 1H); 3.15 (ddd, J = 18.8, 12.3, 1.5 Hz,
1H); 4.14 (s, br, 1H); 4.23 (q, J = 7.1 Hz, 2H); 4.69 (dd, J =
12.3, 5.6 Hz, 1H); 5.72 (s, 1H); 6.87−6.89 (m, 1H); 6.95−7.01
(m, 2H); 7.38−7.44 (m, 2H). 13C NMR (101 MHz, CDCl3) δ
2
14.1, 38.5, 56.2, 62.1, 71.5, 115.3 (d, JFC = 22 Hz), 129.2 (d,
3JFC = 8 Hz), 135.4 (d, 4JFC = 3 Hz), 146.8, 162.5 (d, 1JFC = 246
Hz), 169.2, 170.7.
1
the pooled aqueous layer with MTBE (1.5 L). Assay by H
25
NMR = 99% w/w, ee >99%, [α]D − 128 (c = 1, EtOH).
(R)-2-(4-Fluorophenyl)-2-(trimethylsilyloxy)acetyl
chloride [(R)-3]. A 10 L reactor was charged with (R)-2-(4-
(S)-1-((R)-2-(4-Fluorophenyl)-2-hydroxyacetyl)-4,5-di-
hydro-1H-pyrazole-5-carboxylic acid (8). A 10 L reactor
was charged with acetonitrile (2.5 L), water (55 mL), and
973
dx.doi.org/10.1021/op500134e | Org. Process Res. Dev. 2014, 18, 969−975

Organic Process Research & Development
Article
triethylamine (309 g, 3.06 mol) followed by the addition of
LiBr (531 g, 6.11 mol). Caution: The addition of LiBr is
exothermic! The jacket temperature was set at 25 °C, and (S)-
ethyl 1-((R)-2-(4-fluorophenyl)-2-hydroxyacetyl)-4,5-dihydro-
1H-pyrazole-5-carboxylate 6 (360 g, 1.22 mol) was added at
a reaction temperature of +27 °C. The white suspension was
stirred vigorously (400 rpm) for 6 h. EtOAc (4 L) and 1 M
HCl (aq., 3 L) were added. The aqueous layer (4 L, pH = 1)
was extracted with EtOAc (3 × 1 L). The combined organic
layer was washed with water (1.5 L). The colorless organic
phase was then concentrated. This furnished the titled acid 8 as
a pale yellow solid (315 g, 1.18 mol, 97% yield), dr >98:2, ee
>99%, assay 96% w/w by 1H NMR. HRMS (ESI+) m/e
Hz, 1H); 4.21 (dd, J = 16.0, 5.9 Hz, 1H); 4.55 (dd, J = 12.0, 5.7
Hz, 1H); 5.65 (d, J = 7.0 Hz, 1H); 5.75 (d, J = 7.0 Hz, 1H);
7.10−7.16 (m, 3H); 7.39−7.44 (m, 2H); 7.62−7.72 (m, 3H);
8.71 (t, J = 5.8 Hz, 1H); 9.85 (s, 1H). 13C NMR (101 MHz,
DMSO) δ 38.2, 38.8, 56.3, 70.1, 114.9 (d, 2JFC = 21 Hz), 128.3,
3
128.4, 129.0, 129.1 (d, JFC = 8 Hz), 130.7, 135.6, 136.6, 136.7
(d, 4JFC = 3 Hz), 144.9, 148.4, 161.5 (d, 1JFC = 244 Hz), 169.6,
169.7.
ASSOCIATED CONTENT
■
S
* Supporting Information
1H NMR, 13C NMR, HRMS, and HPLC analyses. Explosivity
screening for compounds 1, 9, and 10. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
267.0775 [(M + H)+, calcd for C12H12FN2O4 267.0781]. H
NMR (600 MHz, DMSO) δ 2.86 (ddd, J = 19.0, 5.4, 1.6 Hz,
1H); 3.23 (ddd, J = 19.0, 12.5, 1.3 Hz, 1H); 4.54 (dd, J = 12.5,
5.4 Hz, 1H); 5.74 (s, 1H); 5.76 (s, br, 1H); 7.08−7.14 (m,
3H); 7.38−7.42 (m, 2H). 13C NMR (101 MHz, DMSO) δ
38.4, 55.5, 70.1, 114.9 (d, 2JFC = 21 Hz), 129.1 (d, 3JFC = 8 Hz),
136.8 (d, JFC = 3 Hz), 148.2, 161.6 (d, JFC = 243 Hz), 169.6,
171.1.
AUTHOR INFORMATION
Corresponding Author
*E-mail: staffan.po.karlsson@astrazeneca.com.
Notes
■
4
1
The authors declare no competing financial interest.
(5-Chloro-2-(1H-tetrazol-1-yl)phenyl)methanamine
(10). Four A 10 L reactor was charged with tert-butyl 5-chloro-
2-(1H-tetrazol-1-yl)benzylcarbamate 9 (240 g, 0.77 mol) and
acetonitrile (1 L) followed by the addition of a 6 M HCl
solution (aq., 1 L, 6.0 mol). The white suspension was stirred
vigorously at 20 °C for 5 h. To the resulting homogeneous
solution was added MTBE (1.5 L) and water (1.5 L). To the
aqueous phase was added EtOAc (3 L) and the mixture was
then cooled to 8 °C. A 10 M aqueous solution of NaOH (1 L)
was slowly added until a pH of 14 was obtained. The organic
layer (4 L) was extracted with water (2 L). The EtOAc solution
(3750 mL) containing the crude titled amine 10 is then used as
such in the next step without further purification. Caution: The
amine 10 is a potential explosive and should not be isolated! A
small sample was concentrated for analytical purposes. An assay
ACKNOWLEDGMENTS
■
We thank our separation science laboratory group for
determination of enantiomeric purities, our analytical specialists
for recording analytical data such as HRMS and HPLC analyses
and Magnus Polla for valuable discussions. We also thank Anna
Pettersen for providing us with X-ray data of AZD8165.
REFERENCES
■
(1) (a) Straub, A.; Roehrig, S.; Hillisch, A. Angew. Chem., Int. Ed.
2011, 50, 4574−4590. (b) Branalt, J.; Gustafsson, D.; Nilsson, I.; Polla,
̊
M. U.S. Patent 0318517, Dec 24, 2009.
(2) Mish, M. R.; Guerra, F. M.; Carreira, E. M. J. Am. Chem. Soc.
1997, 119, 8379−8380.
(3) Kim, H.-O.; Lum, C.; Lee, M. S. Tetrahedron Lett. 1997, 38,
4935−4938.
1
(4) Mini-autoclave: maximum rate of pressure rise 3573 bar/s, onset
temperature 153 °C. Classified as an explosive class 1, rank A. Do not
isolate and handle as a solid.
determination by H NMR gave a strength of 37.6 mg/mL of
1
10 (87% yield). H NMR (600 MHz, DMSO) δ 1.87 (s, br,
2H); 3.50 (s, 2H); 7.56−7.60 (m, 2H); 7.84−7.86 (m, 1H);
9.84 (s, 1H).
(S)-N-(5-Chloro-2-(1H-tetrazol-1-yl)benzyl)-1-((R)-2-(4-
fluorophenyl)-2-hydroxyacetyl)-4,5-dihydro-1H-pyra-
zole-5-carboxamide (1). At 20 °C, a 10 L reactor was
charged with a crude solution of (5-chloro-2-(1H-tetrazol-1-
(5) (a) Nantermet, P. G.; Selnick, H. G.; Barrow, J. C.; Coburn, C.;
Burgey, C.; Robinson, K. A., Lyle, T. A. U.S. Patent 0158218, Aug 21,
2003. (b) Young, M. B.; Barrow, J. C.; Glass, K. L.; Lundell, G. F.;
Newton, C. L.; Pellicore, J. M.; Rittle, K. E.; Selnick, H. G.; Stauffer, K.
J.; Vacca, J. P.; Williams, P. D.; Bohn, D.; Clayton, F. C.; Cook, J. J.;
Krueger, J. A.; Kuo, L. C.; Lewis, D.; Lucas, B. J.; McMasters, D. R.;
Miller-Stein, C.; Pietrak, B. L.; Wallace, A. A.; White, R. B.; Wong, B.;
Yan, Y.; Nantermet, P. G. J. Med. Chem. 2004, 47, 2995−3008.
(6) Mini-autoclave: maximum rate of pressure rise 386 bar/s.
Although at that time, the compound was considered safe to handle as
a solid during normal lab operations, in absence of further testing (i.e.,
UN Time/Pressure and Koenen test), it is recommended to treat the
compound as a potential explosive. Do not heat > 30 °C.
(7) Building block C (R2 = BOC) was outsourced to an external
vendor and in their preparation; the isolation of the highly energetic
primary amine C (R2 = H) was avoided.
yl)phenyl)methanamine 10 (141 g, 0.67 mol) in EtOAc (Vtot
=
3750 mL), (S)-1-((R)-2-(4-fluorophenyl)-2-hydroxyacetyl)-4,5-
dihydro-1H-pyrazole-5-carboxylic acid 8 (179 g, 0.67 mol) and
N-methylmorpholine (102 g, 1.01 mol) followed by the
addition of TBTU (283 g, 0.88 mol) in small portions during
20 min. The product started to precipitate within 5 min of
stirring. The resulting suspension was stirred for 22 h after
which full conversion had been obtained. The mixture was
filtered, and the solid was washed with EtOH (3 × 500 mL)
and water (1500 mL). The product cake was then dried under
reduced pressure furnishing the titled compound 1 as a
colorless solid (182 g, 0.40 mol, 59% yield), assay 98% w/w by
1H NMR, purity >99% by HPLC, ee >99%, dr >99%. An
additional 52 g (0.11 mol, 16% yield) of the titled pure
compound was obtained through recrystallization of material
obtained from the mother liquor. HRMS (ESI+) m/e 458.1142
(8) See for example: (a) Sibi, M. P.; Stanley, L. M.; Jasperse, C. P. J.
Am. Chem. Soc. 2005, 127, 8276−8277. (b) Gerten, A. L.; Slade, M. C.;
Pugh, K. M.; Stanley, L. M. Org. Biomol. Chem. 2013, 11, 7834−7837.
(c) Wang, G.; Liu, X.; Huang, T.; Kuang, Y.; Lin, L.; Feng, X. Org. Lett.
2013, 15, 76−79.
́
(9) See for example: Mendiola, J.; García-Cerrada, S.; Frutos, O.;
Puente, M. L. Org. Process Res. Dev. 2012, 16, 1312−1316. references
cited therein. (b) Xue, Y.-P.; Xu, M.; Chen, H.-S.; Liu, Z.-Q.; Wang,
Y.-J.; Zheng, Y.-G. Org. Process Res. Dev. 2013, 17, 213−220.
(c) Griengl, H.; Osprian, I.; Schoemaker, H.; Reisinger, C.; Schwab,
H. U.S. Patent 0199256, Sep 7, 2006.
1
[(M + H)+, calcd for C20H18N7O3FCl 458.1144]. H NMR
(400 MHz, DMSO) δ 2.70 (ddd, J = 18.9, 5.7, 1.6 Hz, 1H);
3.17 (ddd, J = 18.9, 12.0, 1.4 Hz, 1H); 4.11 (dd, J = 16.0, 5.6
974
dx.doi.org/10.1021/op500134e | Org. Process Res. Dev. 2014, 18, 969−975

Organic Process Research & Development
Article
(10) (a) Hoover, J. R. E.; Dunn, G. L.; Jakas, D. R.; Lam, L. L.;
Taggart, J. J.; Guarini, J. R.; Phillips, L. J. Med. Chem. 1974, 17, 34−41.
(b) Korver, O. Tetrahedron 1970, 26, 5507−5518.
(11) Kelly, S. E.; LaCour, T. G. Synth. Commun. 1992, 22, 859−869.
(12) See for example: (a) Seitaro, K.; Hajime, I.; Kunio, O.; Junya, F.;
Norio, O.; Hiroshi, K.; Yasuhiro, T.; Yoshio, S. Eur. Patent 0373512,
June 20, 1990. (b) Huisgen, V. R.; Koszinowski, J.; Ohta, A.; Schiffer,
R. Angew. Chem. 1980, 92, 198−199. (c) Loevezijn, A.; Venhorst, J.;
Bakker, W. I. I.; Korte, C. G.; Looff, W.; Verhoog, S.; Wees, J.-W.;
Hoeve, M.; Woestijne, R. P.; Neut, M. A. W.; Borst, A. J. M.; Dongen,
M. J. P.; Bruin, N. M. W. J.; Keizer, H. G.; Kruse, C. G. J. Med. Chem.
2011, 54, 7030−7054. (c1) Hiroyuki, Y.; Makoto, O.; Hajime, I.;
Hiroshi, K.; Yoshio, S.; Hiroshi, N. Eur. Patent 0295695, Dec 21, 1988.
(d) Rovnyak, G. C. U.S Patent 4266065, May 5, 1981. (e) Jazouli, M.;
́
Carboni, B.; Carrie, R.; Soufiaoui, M.; Toupet, L. Heteroatom Chem.
1994, 5, 513−518. (f) Curtis, M. D.; Hayes, N. C.; Matson, P. A. J.
Org. Chem. 2006, 71, 5035−5038. (g) Chiericato, M.; Croce, P. D.;
Carganico, G.; Maiorana, S. J. Heterocyclic Chem. 1979, 383−384.
́
(h) Levai, A. J. Heterocyclic Chem. 2002, 39, 1−13.
(13) (a) Simovic, D.; Di, M.; Marks, V.; Chatfield, D. C.; Rein, K. S.
J. Org. Chem. 2007, 72, 650−653. (b) Jun, M. A.; Park, W. S.; Kang, S.
K.; Kim, K. Y.; Kim, K. R.; Rhee, S. D.; Bae, M. A.; Kang, N. S.; Sohn,
S.-K.; Kim, S. G.; Lee, J. O.; Lee, D. H.; Cheon, H. G.; Kim, S. S.; Ahn,
J. H. Eur. J. Med. Chem. 2008, 43, 1889−1902.
(14) The concentrated oily cycloadduct was found to slowly
decompose upon storage at room temperature. Approximately 25%
decomposition was observed after being stored for 20 h.
(15) (a) Elguero, J.; Gonzalez, E.; Jacquier, R. Bull. Chim. Soc. Fr.
1969, 2054−2061. (b) Alkorta, I.; Gonzalez, E.; Jagerovic, N.; Elguero,
J.; Flammang, R. J. Phys. Org. Chem. 2000, 13, 372−381.
(16) Bisprotonation of the intermediate dihydropyrazole is unlikely
as the calculated pKa, using the software “ACD 2012 Expert”, of the
diprotonated species is very low (∼ −9).
(17) We found that one way to epimerize 6/7 was through treatment
with strong bases like DBU. The epimerisation was however not
selective as both of the chiral centres epimerized resulting in
racemization.
(18) Barnhart, R. W.; Dale, D. J.; Ironside, M. D.; Vogt, P. F. Org.
Process Res. Dev. 2009, 13, 1388−1394.
(19) By chiral chromatography analyses of the crude hydrolysis
mixtures and comparison with authentic material of all 4 possible
stereoisomers, we concluded that no epimerisation had occurred at the
CHOH-position of the mandelic acid moiety.
(20) Hogberg, T.; Strom, P.; Ebner, M.; Ramsby, S. J. Org. Chem.
̈
̈
̈
1987, 52, 2033−2036.
(21) Mattsson, S.; Dahlstrom, M.; Karlsson, S. Tetrahedron Lett.
̈
2007, 48, 2497−2499.
(22) The amine 10 was stable for several days as a solution in EtOAc
and only upon prolonged storage; some traces of acetylated amine was
detected.
(23) Wehrstedt, K. D.; Wandrey, P. A.; Heitkamp, D. J. Hazard.
Mater. 2005, A126, 1−7.
(24) 1: DSC: 878 J/g, mini-autoclave: maximum rate of pressure rise
165 bar/s, onset temperature 183 °C, classified as a nonexplosive.
(25) CCDC 997292 contains the supplementary crystallographic
data for AZD8165 supporting the absolute configuration of 1. These
data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/
cif.
(26) Nelson, T. D.; LeBlond, C. R.; Frantz, D. E.; Matty, L.; Mitten,
J. V.; Weaver, D. G.; Moore, J. C.; Kim, J. M.; Boyd, R.; Kim, P.-Y.;
Gbewonyo, K.; Brower, M.; Sturr, M.; McLaughlin, K.; McMasters, D.
R.; Kress, M. H.; McNamara, J. M.; Dolling, U. H. J. Org. Chem. 2004,
69, 3620−3627.
(27) Montalbetti, C. A. G. N.; Falque, V. Tetrahedron 2005, 61,
10827−10852.
975
dx.doi.org/10.1021/op500134e | Org. Process Res. Dev. 2014, 18, 969−975