Development and Scale-Up of an Improved Manufacturing Route to the ATR Inhibitor Ceralasertib

Article abstract of DOI:10.1021/acs.oprd.0c00482

Ceralasertib is currently being evaluated in multiple phase I/II clinical trials for the treatment of cancer. Its structure, comprising a pyrimidine core decorated with a chiral morpholine, a cyclopropyl sulfoximine and an azaindole, makes it a challenging molecule to synthesize on a large scale. Several features of the medicinal chemistry and early development route make it unsuitable for the long-term commercial manufacture of the active pharmaceutical ingredient. We describe the investigation and development of a new and improved route which introduces the cyclopropyl moiety in a novel process from methyl 2,4-dibromobutyrate. Following construction of the pyrimidine ring, large-scale chlorination with phosphoryl chloride was performed with a safe and robust work-up. An SNAr reaction required an innovative work-up to remove the unwanted regio-isomer, and then a Baeyer-Villiger monooxygenase enzyme was used to enable asymmetric sulfur oxidation to a sulfoxide. A safe and scalable metal-free sulfoximine formation was developed, and then optimization of a Suzuki reaction enabled the manufacture of high-quality ceralasertib with excellent control of impurities and an overall yield of 16%.

Full text of DOI:10.1021/acs.oprd.0c00482

pubs.acs.org/OPRD
Article
Development and Scale-Up of an Improved Manufacturing Route to
the ATR Inhibitor Ceralasertib
Mark A. Graham,* Hannah Askey, Andrew D. Campbell, Lai Chan, Katie G. Cooper, Zhaoshan Cui,
Andrew Dalgleish, David Dave, Gareth Ensor, Rita Galan, Peter Hamilton, Claire Heffernan,
Lucinda V. Jackson, Dajiang Jing, Martin F. Jones, Pengpeng Liu, Keith R. Mulholland,
Mohammed Pervez, Michael Popadynec, Emma Randles, Simone Tomasi, and Shenghua Wang
Cite This: https://dx.doi.org/10.1021/acs.oprd.0c00482
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Ceralasertib is currently being evaluated in multiple phase I/II clinical trials for the treatment of cancer. Its structure,
comprising a pyrimidine core decorated with a chiral morpholine, a cyclopropyl sulfoximine and an azaindole, makes it a challenging
molecule to synthesize on a large scale. Several features of the medicinal chemistry and early development route make it unsuitable
for the long-term commercial manufacture of the active pharmaceutical ingredient. We describe the investigation and development
of a new and improved route which introduces the cyclopropyl moiety in a novel process from methyl 2,4-dibromobutyrate.
Following construction of the pyrimidine ring, large-scale chlorination with phosphoryl chloride was performed with a safe and
robust work-up. An SNAr reaction required an innovative work-up to remove the unwanted regio-isomer, and then a Baeyer−Villiger
monooxygenase enzyme was used to enable asymmetric sulfur oxidation to a sulfoxide. A safe and scalable metal-free sulfoximine
formation was developed, and then optimization of a Suzuki reaction enabled the manufacture of high-quality ceralasertib with
excellent control of impurities and an overall yield of 16%.
KEYWORDS: ceralasertib, AZD6738, ATR inhibitor, cyclopropanation, phosporyl chloride, SNAr, BVMO, sulfoximine, Suzuki−Miyaura,
Suzuki
processing times. We realized that it would be desirable to
undergo some directed evolution of the enzyme to improve the
efficiency of this process but decided that it would be prudent to
establish the commercial route before this undertaking.
Rhodium was used in the formation of sulfoximine 9, and we
were concerned about this precious metal from the cost and
sustainability perspective. However, the major issue with this
route centered around the cyclopropanation of sulfoximine 10.
This cyclopropanation process involves α-deprotonation of the
sulfoximine and reaction with 1-bromo-2-chloroethane. This
was a low-yielding process because of the propensity of the
product to over-react with 1-bromo-2-chloroethane to give the
N-alkylated impurity 12 as a byproduct. To minimize this issue,
control of reaction time was critical, and a continuous stirred
tank process was developed to achieve a moderate yield of 51%.
Furthermore, impurity 12, 1-bromo-2-chloroethane, and other
related impurities are potential mutagenic substances which we
ideally want to avoid late in the synthetic route. We believed that
this was an undesirable process for the potential future
commercial manufacture of drug substance.
INTRODUCTION
■
The kinase ataxia telangiectasia mutated and Rad3 related
(ATR) is a key regulator of the DNA-damage response of cells,
and inhibition of this repair pathway can lead to cell death.
Ceralasertib (AZD6738, 1) was identified by AstraZeneca as a
potent and selective inhibitor of ATR1,2 and is currently being
evaluated in multiple phase I/II clinical trials for the treatment of
cancer. Its structure, comprising a pyrimidine core decorated
with a chiral morpholine, a cyclopropyl sulfoximine and an
azaindole, makes it a challenging molecule to synthesize on a
large scale. The medicinal chemistry route to ceralasertib was
adapted and developed for early clinical active pharmaceutical
ingredient (API) manufacturing campaigns (Scheme 1).3
However, there were several limitations and issues with this
synthetic route, and we felt that a new route would be required
to provide efficient manufacturing processes suitable for future
commercial drug substance.
The early development route was long, with a total step count
of 14 steps from the chiral building block D-alaninol when we
include the 3 steps that were used to manufacture (R)-3-
methylmorpholine (2) using the route described by Mahajan4
and the single step used to manufacture boronate ester 14. There
was also an undesirable linear step count of 13. The expensive
chiral morpholine 2 was introduced at the start of the route, so it
would be preferential from a cost perspective to introduce this at
a later stage. The enzymatic S-oxidation to form sulfoxide 8 was
not fully optimized and required high process volumes and
Received: October 27, 2020
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 1. Early Development Route to Ceralasertib
Although this aldehyde was unsuitable for our planned
synthetic route, it opened the possibility of synthesizing a related
ester or acid which could potentially be useful. Methyl 2,4-
dibromobutyrate (17)6 is commercially available in bulk
quantities, and we surmised that this could react with sodium
thiomethoxide in a similar manner to yield the analogous ester.
Gratifyingly, sodium thiomethoxide reacted selectively at C2 of
the ester, and then treatment with 1 equiv of sodium hydroxide
resulted in cyclization to form ester 18 (Scheme 3). Further
addition of excess sodium hydroxide resulted in ester
saponification to yield acid 19 in 86% crude yield following
acidic work-up.
We proceeded to develop this process by changing the solvent
from tetrahydrofuran (THF) to 2-methyltetrahydrofuran (2-
MeTHF). The use of 2-MeTHF as the main solvent enabled us
to efficiently telescope the process without the need for an
additional solvent such as ethyl acetate for the work-up. The use
of methanol as a cosolvent helped to improve the solubility of
sodium thiomethoxide and increased the rate of reaction. After
an investigation of bases (Table 1), we established that the use of
sodium methoxide resulted in more efficient deprotonation and
cyclization to the form the cyclopropyl ring. We continued to
use aqueous sodium hydroxide for the ester hydrolysis.
Following work-up, acid 19 was obtained as a dry 2-MeTHF
solution in 93% yield on 118 kg scale (Scheme 4).
RESULTS AND DISCUSSION
■
Initial Route Scouting. As part of a route design exercise,
the desire to avoid this unfavorable cyclopropanation process led
us to consider alternative options to introduce the cyclopropyl
moiety. A thorough literature search confirmed that are no
known examples of direct C−C bond forming reactions between
a thiocyclopropyl synthon and a pyrimidine ring system. We
therefore decided to investigate alternative ways of synthesizing
the thiomethylcyclopropyl moiety, and we were interested in
building the pyrimidine ring via classical condensation
chemistry. A literature investigation identified the work of
Kanel and Dingwall,5 which demonstrated the ring opening of 3-
chloro-2-methoxytetrahydrofuran (14) with thionyl chloride
and pyridine to give 2,4-dichlorobutanal, which was isolated as
sodium bisulfite adduct 15 (Scheme 2). This was then reacted
with sodium thiomethoxide followed by sodium hydroxide to
give aldehyde 16 in 80% yield.
Scheme 2. Kanel Synthesis of 1-
(Methylsulfanyl)cyclopropanecarbaldehyde
To enable the construction of a pyrimidone ring, we needed to
convert acid 19 to β-keto ester 20, which could subsequently be
Scheme 3. Initial Investigation into Formation of Acid 19
B
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
a
sulfoximine formation. It is clear that these reaction conditions
do not tolerate the presence of electron-rich heteroaromatic
compounds.
Table 1. Investigation of Bases for Conversion of 17 to 18
entry
base
time for full conversion to 18
isolated yield of 19
1
2
3
NaOH
NaOMe
Et3N
24 h
1 h
no conversion
72%
83%
0
Development of the Dichloropyrimidine Route Strat-
egy. We therefore needed to revise our route strategy to an
alternative route where the sulfoximine is formed before the
introduction of the azaindole ring. We decided that the best way
to do this would be to use 2-chloropyrimidine 13 as the final
isolated intermediate. This would allow the introduction of the
azaindole via Suzuki chemistry in a way similar to the early
development route. Attempted reaction of β-ketoester 20 with
urea resulted in only a trace amount of pyrimidone 31, so we
switched our attention to the more reactive thiourea which
formed thiopyrimidone 27. Initially, sodium ethoxide was
evaluated as a base for this reaction based on literature
precedent,13 but we observed a moderate yield of only 60%,
mainly due to the formation of unwanted impurities as a result of
competing side reactions. The three most significant impurities
were tentatively assigned based on liquid chromatography mass
spectrometry data as 28 and 29 (formed by reaction of thiourea
with 2 molecules of 20) and 30 (formed as a result of Claisen
ester condensation). An investigation of alternative bases and
solvents (Table 2) showed that varying these parameters had a
significant impact on the ratio of observed products and
identified potassium tert-butoxide (entry 2) as a superior base
for favoring the desired product, while minimizing impurity
formation.
The increased steric bulk of tert-butoxide compared to
ethoxide could decrease the rate of deprotonation of β-ketoester
20, hence suppressing the formation of the Claisen ester
condensation product 30, but intriguingly, ethanol proved to be
a more efficient solvent than tert-butanol (entry 8) for this
transformation. The use of potassium hydroxide (entry 3) and
potassium carbonate (entry 4) led to significant amounts of
impurity 28, and the use of NaOMe resulted in significant
amounts of Claisen impurity 30. The use of potassium tert-
butoxide (entry 2) was advantageous over sodium tert-butoxide
(entry 7), so there also appears to be a counterion effect. We
were able to develop an efficient process using potassium tert-
butoxide in ethanol, following a solvent swap from 2-MeTHF. A
a
Reactions performed in 2-MeTHF (4 rel vol) and MeOH (2 rel vol)
after the addition of NaSMe (1.1 equiv) at 20 °C.
condensed with an amidine derivative. We applied the well-
known methodology for conversion of carboxylic acids to β-keto
esters which was reported by Challenger et al. on a multikilo-
gram scale.7 This involved acylation of ethyl potassium malonate
with the imidazolide formed by reacting acid 19 with
carbonyldiimidazole, followed by decarboxylation under mild
conditions in the presence of triethylamine and magnesium
chloride. We modified the reported conditions to use 2-MeTHF
as the solvent, which enabled us to telescope acid 19 directly into
this process as a 2-MeTHF solution (Scheme 4). Our modified
process afforded β-keto ester 20 as a 2-MeTHF solution in 90%
yield.
Amidine 22 was prepared from 7-azaindole-4-carbonitrile
(21) by reaction with hydroxylamine, followed by hydro-
genation in the presence of acetic anhydride in a process similar
to that of Dorsey (Scheme 4).8 Condensation of amidine 22
with β-keto ester 20 proceeded efficiently in the presence of
sodium ethoxide to yield pyrimidone 23, and then chlorination
of the pyrimidone was achieved with phosphoryl chloride. (R)-
3-Methylmorpholine hydrochloride reacted with chloropyrimi-
dine 24 via SNAr; then, sulfur oxidation was achieved on small
scale using Kagan conditions9 to yield sulfoxide 26 with only
moderate stereoselectivity. However, attempts to convert this
sulfoxide to sulfoximine 1 (ceralasertib) proved to be
unsuccessful. Rhodium-catalyzed conditions that were used in
the early development route were initially attempted;10 then,
metal-free sulfoximine formation conditions recently developed
by Bull11,12 also proved to be unsuccessful. This is not only
consistent with the general literature but also correlates well with
the work of Bull, which showed that spiking reaction mixtures
with electron-rich heterocycles shut down the desired
Scheme 4. Initial New Route Investigation Using Amidine
C
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 2. Investigation of Bases and Solvents for Formation of Thiopyrimidone 27
HPLC area %
entry
base
solvent
product 27
SM 20
impurity 28
impurity 29
impurity 30
isolated yield
1
2
3
4
5
6
7
8
NaOEt
KOtBu
KOH
K2CO3
NaOMe
NaOEt
NaOtBu
KOtBu
EtOH
EtOH
EtOH
EtOH
EtOH
n-BuOH
EtOH
t-BuOH
81.5
93.1
16.7
22.6
26.5
75.9
79.0
6.1
1.5
0.7
2.5
21.2
0.7
2.2
1.5
67
50.3
4.7
2.4
0
3.5
2.7
8.4
18.1
4.9
12.8
8.4
NA
0.3
0
60
85
2.2
1.5
32.0
1.1
0.5
NA
NA
NA
NA
NA
NA
58
0.6
1.1
a
86.0
NA
a
Reaction time of 24 h.
Scheme 5. Optimized New Route to Ceralasertib
facile work-up was developed involving acidification to pH 1
with hydrochloric acid resulting in the crystallization of
thiopyrimidone 27, affording a yield of 78% upon scale-up to
123 kg on a pilot plant (Scheme 5).
Conversion of thiopyrimidone 27 to pyrimidone 31 was
efficiently achieved using a well-known procedure with chloro-
acetic acid using water as the solvent.14 Full conversion to
product was achieved by heating to 95 °C for 4 h, and then
crystallization conveniently occurred upon cooling to 20 °C to
afford pyrimidone 31 in 83% yield. This was a remarkably
straightforward, green, and highly efficient process which did not
require an organic solvent.
Process Development of Chlorination and SNAr
Reaction. To enable the introduction of (R)-3-methylmorpho-
line and azaindole, conversion of pyrimidone 31 to dichlor-
opyrimidine 32 was required. This was achieved using
phosphoryl chloride with N,N-diethylphenylamine, as reported
by Frutos.15 (Scheme 6) Initial investigations using solvents
such as toluene and acetonitrile gave a slow rate of reaction
resulting in long reaction times and formation of many
impurities. A suitable process was developed using a mixture
D
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 6. Chlorination of Pyrimidone Followed by Quench
Table 3. Optimization of Parameters for SNAr Reaction
solvent
base
DIPEA
DIPEA
DIPEA
DIPEA
temp (°C)
34 (HPLC area %)
35 (HPLC area %)
reaction time
MeCN
MeCN
toluene
MeOH
DMSO
DMSO
50
20
50
20
20
20
77
82
67
83
80
83
23
18
33
17
20
17
complete in 24 h
>3 days
>3 days
complete in 24 h
>3 days
complete in 18 h
a
DIPEA
Et3N + K2CO3
a
12% OMe SNAr product observed.
of phosphoryl chloride and N,N-diethylphenylamine without
additional solvent. Good conversion to dichloropyrimidine 32
was observed over 6 h at 95 °C. However, there were some
significant issues encountered with the quench. Quenching
directly into water resulted in precipitation of the product.
Initially, this looked to be a viable process, but we realized that it
was important to perform the quench at temperatures in the
region of 20−30 °C in the presence of a base because the
formation of metastable partially hydrolyzed intermediates
during a cold quench into water could result in significant
hazards upon scale-up of the process.16 A stress test of the
quench mixture using 25% NaOH at 50 °C over 15 h resulted in
significant hydrolysis back to pyrimidones 33 (51%) and 31
(24%). This issue was overcome by quenching the reaction
mixture into a solution of aqueous sodium acetate, with
simultaneous addition of aqueous NaOH to maintain the pH
in the range of 6−7. Dichloropyrimidine 32 was stable in
aqueous conditions at this pH range, and the resulting
precipitated solid could be collected by filtration (Scheme 6).
This process afforded dichloropyrimidine 32 in a good yield of
79% and high-performance liquid chromatography (HPLC)
purity of 95% at 51 kg scale. However, the solid was
contaminated with a black tar which was problematic for
crystallization in the following stage. This was overcome by
extracting the crude product from the quenched mixture into n-
heptane, followed by treatment with activated carbon. Solvent
swap from heptane to dimethyl sulfoxide (DMSO) could then
be performed before the next stage. The optimized process
afforded 32 as a DMSO solution in 84% yield as measured by
NMR assay.
SNAr chemistry was required to introduce the chiral (R)-3-
methylmorpholine moiety to the 4-position of pyrimidine 32.
SNAr reaction under basic conditions clearly favored formation
of 4-morpholino isomer 34, but a significant amount of 2-
morpholino isomer 35 also formed. A screen of various solvents,
bases, and temperatures demonstrated that the choice of the
solvent had a significant impact on the ratio of 34 to 35 (Table
3). We chose to use (R)-3-methylmorpholine hydrochloride
because of the substance being a solid, which is easy to handle
and commercially available in bulk. Our primary objective was
improving the selectivity for 4-morpholino isomer 34, and it was
clear that lower temperatures tended to favor increased amounts
of 34. However, the rate of reaction was an issue, particularly at
lower temperatures. Methanol as a solvent increased the rate of
reaction and gave improved selectivity, but the SNAr product
with MeOH was also observed. DMSO appeared to be a suitable
solvent, but the reaction rate was slow when N,N-diisopropy-
lethylamine (DIPEA) was used alone as a base. However, the
combination of both a tertiary amine and potassium carbonate
significantly improved the rate of reaction and gave the best
observed ratio of 83:17. DIPEA and triethylamine gave similar
results, but triethylamine was easier to remove in an aqueous
work-up, so we decided to proceed with the combination of
triethylamine and potassium carbonate.
For this to become a viable process, removal of the unwanted
2-morpholino isomer 35 was necessary. The crude product as a
mixture of isomers was isolated as a gum, and we were unable to
develop crystallization as the free base or salt. The problem was
solved by considering the difference in pKa and log P between
the two isomers. The desired isomer 34 is both more basic than
35 (calculated pKa of 4.9 compared to 3.8) and more
E
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 4. Partitioning of Pyrimidine Isomers between MTBE and Aqueous Solutions
aqueous layer
MTBE layer
aqueous solution
34 (HPLC area %)
35 (HPLC area %)
34 (HPLC area %)
35 (HPLC area %)
pH 2.5 buffer
0
100
100
98
0
0
0
2
11
83
73
40
5
17
27
60
95
100
2 M HCl solution
4 M HCl solution
6 M HCl solution
12 M HCl solution
89
0
Figure 1. log P of 34 vs Δlog P for a range of organic solvent/water systems using COSMOtherm. Symbol color and size represent ICH class and
boiling point, as detailed in the legend.
hydrophilic (calculated log P of 2.0 compared to 4.6).17 We
hypothesized that this could work in our favor when performing
an extraction into aqueous acid. To investigate this hypothesis,
the crude isolated product was dissolved in methyl tert-butyl
ether (MTBE), and then samples were washed with a range of
acidic aqueous solutions to investigate partitioning of the two
isomers between the organic and aqueous layers (Table 4). With
pH 2.5 buffer solution, both isomers remained fully in the
organic layer. However, when the organic solution was washed
with 2 M hydrochloric acid, it became apparent that some of 4-
morpholino isomer 34 was extracted into the aqueous layer. The
use of 4 M hydrochloric acid resulted in most of 34 being
extracted into the aqueous layer, but all of 2-morpholino isomer
35 remained in the organic layer. Higher concentrations of
hydrochloric acid resulted in all of 34 together with some 35
being extracted into the aqueous layer.
With this encouraging result in hand, we then performed
some predictive modelling of partition coefficients to decide on
the optimal organic solvent for this extraction. COSMOtherm
software18−21 was used to calculate and plot the log P of 4-
morpholino isomer 34 against the Δlog P between isomers 34
and 35 (Figure 1).
We selected the set of solvents for log P predictions from our
internal solvent database based on a combination of solvent
properties, following an in-house developed approach.22 We
disregarded any ICH class 1 solvents and solvents with boiling
points greater than 200 °C. Ideally, we would choose a solvent
with a low log P and a high Δlog P to facilitate extraction of 4-
morpholino isomer 34 into the aqueous layer while leaving 2-
F
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
morpholino isomer 35 in the organic layer. MTBE would be a
preferable solvent to toluene because 34 has a lower log P in this
solvent, but Δlog P has a similar value. However, the alkane class
of solvent provides the best solution, being best for extraction
efficiency (low log P of 34) and selectivity (highest Δlog P
between 34 and 35). We chose heptane as a process-friendly
solvent which provided optimal partitioning for our needs.
n-Heptane also provided the ideal solvent for work-up of the
SNAr reaction in DMSO. At the end of the reaction, water and n-
heptane were charged to DMSO solution, resulting in a clean
partition with both product isomers 34 and 35 in the organic
layer. DMSO, excess base, and salts were observed in the
aqueous layer with only a trace of the product. A single
extraction of the n-heptane solution with 4 M aqueous
hydrochloric acid resulted in extraction of >98% of 4-
morpholino isomer 34 into the aqueous phase, leaving 35 in
the organic layer. On scale-up, an additional extraction with 4 M
hydrochloric acid was performed to ensure optimal yield. The
acidic extracts were then neutralized with NaOH and extracted
into MTBE. Distillative solvent swap to propan-2-ol was easily
achieved, followed by crystallization to isolate pure 34. This
process enabled the pilot plant production of 34 as a single regio-
isomer in 99.8% purity on 77 kg scale.
Baeyer−Villiger Monooxygenase-Catalyzed Asym-
metric Sulfur Oxidation. For the early development route,
biocatalytic sulfur oxidation was performed on a different
substrate using a Baeyer−Villiger monooxygenase (BVMO)
enzyme to afford the sulfoxide in 79% yield and a chiral purity of
>99 de.10 We were keen to build on this successful development
to install the chiral sulfoxide and subsequent sulfoximine for this
new route and identified a suitable BVMO enzyme23 which
achieved excellent stereo control to give sulfoxide 36 with >99%
de. The enzyme uses oxygen, together with reduced
nicotinamide adenine dinucleotide phosphate (NADPH) as a
cofactor, which is oxidized to NADP+. Recycling of NADP+ to
NADPH is achieved with an additional keto-reductase (KRED)
enzyme which oxidizes the cosolvent isopropanol to acetone
(Scheme 7). The reaction was performed at pH 9 using a
using n-heptane as an antisolvent to afford sulfoxide 36 in 88%
yield and >99% de on 64 kg scale.
Process Development of Metal-Free Sulfoximine
Formation. The early development route used a rhodium
acetate catalyzed process with diacetoxyiodobenzene as the
oxidant and trifluoroacetamide as the nitrogen source, which
required subsequent hydrolysis of the acetamide. The cost of
rhodium acetate was a concern for a commercial manufacturing
process, and we were interested in a recent report by Bull which
described metal-free conditions for sulfoximine formation.11 We
initially applied the reported conditions to sulfoxide 36 using
diacetoxyiodobenzene (3 equiv) and ammonium carbamate (4
equiv) as a nitrogen source in methanol at 20 °C. Good
conversion to product 13 was observed over the course of 1 h at
20 °C, but we were mindful of the potential to generate explosive
iodosobenzene in this process, so a moderate reaction
temperature and a safe extractive work-up was required which
did not involve heating or distillation. Toluene was also
investigated as a potential solvent, but the rate of reaction
proved to be slower, with only 33% conversion to 13 over 23 h at
20 °C. A mixture of toluene (8 relative volumes) and methanol
(2 relative volumes) resulted in a suitable rate of reaction at 20
°C and allowed for an extractive aqueous work-up (Scheme 5).
The work-up involved extraction of the product into 30% citric
acid solution, followed by basification of the aqueous layer with
sodium hydroxide and then extraction into a 3:1 mixture of ethyl
acetate and THF. Distillative solvent swap to isopropanol was
performed, and then a solution of hydrogen chloride in
isopropanol was added to form the hydrochloride salt of 13,
which was isolated following the addition of MTBE as an
antisolvent. A small amount of oxidation impurity (up to 3%)
was observed, which was tentatively assigned as dihydro-1,4-
oxazine 37, based on MS fragmentation data. The subsequent
oxidation impurity in the API could not be efficiently purged by
crystallization, so we needed to control its formation at this
stage. We developed this process to control the impurity to
<0.2% by decreasing the diacetoxyiodobenzene charge from 3.0
to 2.1 equiv, decreasing the reaction temperature to 5 °C, and
completing the stage by performing a slurry of the hydrochloride
salt of 13 in methanol. The optimized process afforded the
hydrochloride salt of sulfoximine 13 in 83% yield on 30 kg scale
at >99% purity.
Scheme 7. Asymmetric Sulfur Oxidation with BVMO Enzyme
Process Development of Suzuki Reaction. In common
with the early development route, the final stage involved a
Suzuki coupling between chloropyrimidine 13 and boronate
ester 14. First, we were keen to improve the process which was
used to manufacture the boronate ester from the commercially
available aryl bromide 38 (Scheme 8). The early development
route used a Miyaura borylation process which used B2Pin2 with
tetrakis(triphenylphosphine)palladium(0) as the catalyst and
potassium acetate as the base in 1,4-dioxane. Although
conversion to boronate ester 14 proceeded well under these
conditions, the work-up was inefficient involving filtration,
dipotassium hydrogen phosphate buffer. As was the case for the
previous biocatalytic process, the key engineering challenge was
the mass transfer of sufficient oxygen, and this was achieved by
sparging a mixture of air and nitrogen (1:2) through the reaction
mixture at a high rate of stirring. After full conversion to
sulfoxide was achieved, the enzyme was denatured using
hydrochloric acid and sodium chloride. The solids were
removed by filtration, and then the filtrate was extracted with
ethyl acetate. The product was crystallized from ethyl acetate
Scheme 8. Optimized Miyaura Borylation Reaction
G
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 5. Investigation of Catalysts and Bases in 1-BuOH for Suzuki Coupling
HPLC area % after 23 h
entry
catalyst
base
equiv of 14
1
40
41
1
2
3
4
5
6
7
8
Pd(dppf)Cl2
Pd(dppf)Cl2
Pd(dppf)Cl2
Pd(dppf)Cl2
Pd(PPh3)2Cl2
Pd(PPh3)2Cl2
Pd(PPh3)2Cl2
Pd(PPh3)2Cl2
K2CO3
K2CO3
K3PO4
K3PO4
K2CO3
K2CO3
K3PO4
K3PO4
1.1
1.3
1.1
1.3
1.1
1.3
1.1
1.3
75.6
91.3
89.0
96.3
92.4
94.9
96.7
94.2
16.2
3.5
6.5
0.8
0.9
0.2
0.2
0.7
0.6
0.4
0.1
0.4
0.6
0.6
0.4
0.6
Table 6. Investigation of Catalysts, Bases, and Solvents over 2 h for Suzuki Coupling
HPLC area % after 2 h
entry
catalyst
base
solvent
1
40
41
other imps >0.2%
1
2
3
4
5
6
7
8
Pd(PPh3)2Cl2
Pd(PPh3)2Cl2
Pd(PPh3)2Cl2
Pd(dppf)Cl2
Pd(dppf)Cl2
Pd(PPh3)2Cl2
Pd(dppf)Cl2
Pd(PPh3)2Cl2
Pd(dppf)Cl2
Pd(dppf)Cl2
Pd(PPh3)2Cl2
K2CO3
K3PO4
K2CO3
K2CO3
K3PO4
K2CO3
K2CO3
K3PO4
K2CO3
K3PO4
K3PO4
2-MeTHF
2-MeTHF
iPrAc
ethanol
iPrAc
ethanol
iPrAc
ethanol
2-MeTHF
2-MeTHF
ethanol
29.2
30.22
32.4
55.24
61
62.02
71.4
78.54
80.9
85.69
85.80
0.03
0.13
0.06
0.11
0.48
0.13
0.24
0.08
0.84
0.56
0.1
0.32
0.18
0.4
0.53
0.2
1.09
0.25
0.53
0.23
0.17
0.36
1
1
1
4
2
2
4
2
5
4
4
9
10
11
distillation, and then uncontrolled precipitation by addition of
water, followed by slurrying with MTBE. We were unsure
whether the air-sensitive catalyst tetrakis(triphenylphosphine)-
palladium(0) was optimal and were uncomfortable about the
use of ICH class 2 solvent 1,4-dioxane in a penultimate stage
before the API. A substantial screen was performed using high-
throughput experimentation in a 96-well plate format in a
glovebox. The main aim of this screen was to find a catalyst, a
base, and a solvent system capable of delivering the clean
product with the required speed and enabling a smooth time-
efficient work-up process. The screen24 identified several
potential catalysts, but there was a clear trend in reaction
rates. The reaction rate declined in the order Pd(PPh3)2Cl2 ≈
Pd(OAc)2/PPh3 > Pd(dba)2/PPh3 > Pd(PPh3)4. A key
byproduct that we wanted to avoid was the proto-deboronated
compound 39. The amount of 39 followed the general trend
Pd(dba)2/PPh3 ≥ Pd(PPh3)4 > Pd(OAc)2/PPh3 ≈ Pd-
(PPh3)2Cl2. Pd(PPh3)2Cl2 was clearly the preferred catalyst
for both the reaction rate and avoidance of 39, and 2 mol %
catalyst worked well for this process. The choice of base clearly
impacted the rate of reaction, giving a reaction rate trend of
NMe4OAc > CsOAc > KOAc. However, the amount of impurity
39 followed the same trend. We decided that suppression of
impurity formation was the most important factor, so decided to
use KOAc as the base for this reaction. In terms of solvent, 2-
MeTHF and iPrOAc offered a good balance of properties for
both the reaction and work-up. Boronate ester 14 had a lower
solubility in iPrOAc, resulting in higher yields from the final
crystallization process, so this was chosen as the reaction solvent.
Following the reaction, an aqueous work-up was performed to
remove potassium salts. Further aqueous washes were necessary
to remove boron-related impurities. Treatment with mercap-
toalkyl silica was performed to scavenge excess palladium to less
than 100 ppm, which ensured that good control of impurities
H
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
was achieved in the subsequent Suzuki stage. The iPrOAc
solution was azeotropically dried by distillation, and then
crystallization was achieved using MTBE as the antisolvent. A
yield of 75% was achieved on scale-up to pilot plant using 41 kg
of aryl bromide 38 to afford boronate ester 14 with less than
0.1% of the debrominated byproduct 39.
Table 7. Investigation of Aqueous Solutions for Scavenging
Pd from 2-MeTHF Solution
molar equiv
of scavenger
wrt Pd
Pd concentration
after aqueous wash
a
entry
scavenger
L-cysteine
additive
none
DIPEA
DIPEA
(ppm)
1
2
3
50
50
50
28
1
28
The final Suzuki coupling between chloropyrimidine 13 and
boronate ester 14 had previously been well developed for the
early development route. Although this process was high
yielding, we were concerned about the complexity and number
of unit operations required for the work-up. The previous
process used Pd(dppf)Cl2 as the catalyst, potassium carbonate
as the base, and ethanol as the reaction solvent. The work-up
involved a solvent swap from ethanol to ethyl acetate, followed
by aqueous wash and then treatment with a solid-supported
silica thiol to scavenge palladium. Finally, another solvent swap
to 1-butanol was required, followed by crystallization. We
initially evaluated 1-butanol as a single solvent for reaction,
work-up, and crystallization (Table 5). Although the reaction
proceeded well, we observed the formation of dimeric impurity
40 which was tentatively assigned based on MS data. Dimeric
impurity 40 was particularly prevalent when only 1.1 equiv of
boronate ester 14 was used in the reaction with Pd(dppf)Cl2 as
the catalyst (Table 5, entries 1 and 3). Although the use of
Pd(PPh3)Cl2 as the catalyst appeared to give a cleaner reaction
profile (Table 5, entries 5−8), the API was contaminated with
residual boronate ester 14 following crystallization. We focused
on ethanol as a solvent with a single solvent swap to a suitable
solvent, which was immiscible with water for work-up and
suitable for crystallization. Although 2-MeTHF gave an
unfavorable profile as a reaction solvent, it offered an ideal
solvent for aqueous work-up and crystallization. In particular, we
were able to purge the homodimer impurity 41 during the
crystallization. The use of Pd(dppf)Cl2 carried a higher risk of
forming impurity 40 and other unknown impurities (Table 6,
entries 4, 5, 7, 9, and 10), which could not be purged by
crystallization, so Pd(PPh3)2Cl2 (1.3 mol %) was chosen as the
catalyst. Potassium phosphate was chosen as the base because it
offered a faster rate of reaction than potassium carbonate. At the
end of reaction, solvent swap to 2-MeTHF was performed.
Overall, this allowed us to improve the work-up, palladium
scavenging, and crystallization process.
An investigation of palladium scavenging agents identified L-
cysteine as an efficient scavenger which could be used as an
aqueous solution to wash the 2-MeTHF solution. Further
investigation of additives in combination with scavengers in
aqueous solution suggested that fully deprotonated L-cysteine is
more effective at scavenging palladium (Table 7, entries 2, 5, and
6). Although DIPEA worked well in combination with L-
cysteine, inorganic bases were preferred because of better
aqueous solubility. The level of palladium observed in the
organic solution decreased from 308 to 2 ppm when washed
with the preferred combination of aqueous L-cysteine with
potassium phosphate (Table 7, entry 6).
Despite the successful palladium scavenging process, we were
mindful of the fact that the final bond-forming stage was a
palladium-catalyzed Suzuki process, and there was a risk of
palladium contamination in the API. To mitigate this risk and
ensure that other impurities were purged to acceptable levels
while the API was isolated in the desired crystalline form and
particle size distribution, we decided to develop a final
recrystallization. We identified isopropanol and water as a
binary solvent system, which allowed full dissolution of the API
L-cysteine
L-cysteine ethyl
ester
4
5
6
7
L-cysteine
L-cysteine
L-cysteine
150
150
150
20
HCl
20
4
2
NaOH
K3PO4
none
trithiocyanuric
acid
30
8
trithiocyanuric
acid
20
activated
carbon on
Celite
65
Quadrasil-MP
150
none
10
a
Pd concentration measured as 308 ppm before aqueous wash.
(1) in 4 relative volumes of the solvent at 60 °C. However,
cooling crystallization and isolation of the API using this binary
solvent system resulted in disappointing yields. To achieve an
acceptable crystallization recovery of >85%, azeotropic removal
of water by reduced pressure distillation was required to ensure
that the water content was ≤0.6% w/w by Karl Fischer analysis.
We also knew that 1-butanol could be used for crystallization of
1 from the work performed in the early development program,3
but the isopropanol distillation offered the additional advantage
of purging impurity 41. Our final recrystallization process
involved dissolving the crude solid in 2.5 relative volumes of
isopropanol and 1.2 relative volumes of water. This solution was
filtered to remove particulate impurities, and then azeotropic
drying of the solution by reduced pressure distillation with the
addition of further isopropanol was performed. When the water
content was ≤0.6% by Karl Fischer analysis, seeding and cooling
to 5 °C enabled efficient crystallization. This recrystallization
was a key part of the overall process, resulting in high-quality API
with very good control of key impurities. Purging of the Ames-
positive boronate ester 14 was critical, and all isolated batches of
API contained <0.01% of 14 or the corresponding boronic acid,
as confirmed by residual boron analysis. Residual palladium was
also well controlled to less than 3 ppm in all batches. This
combined Suzuki reaction and recrystallization process afforded
high-quality ceralasertib in 70% yield on a scale of 32 kg.
CONCLUSIONS
■
In conclusion, we have developed a new and improved route to
ceralasertib, which solves many of the scale-up issues faced with
the early development route. A low-yielding cyclopropanation
which generated mutagenic impurities has been replaced by a
novel and high-yielding cyclopropanation from methyl 2,4-
dibromobutyrate. Following construction of the pyrimidine ring,
large-scale chlorination with phosphoryl chloride has been
developed which enables safe quenching of phosphoryl chloride
without product decomposition. An innovative work-up of SNAr
reaction has been developed to remove the unwanted regio-
isomer, based on differences in pKa and partition coefficients.
High-yielding sulfur oxidation was developed using a BVMO
enzyme, and then a safe and scalable metal-free sulfoximine
formation was developed, followed by an optimized Suzuki
reaction. The chiral morpholine 2 was manufactured in three
steps from D-alaninol using the route described by Mahajan4 in
I
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
an overall yield of 32%. When we consider the manufacture of 2
as part of the longest linear sequence for the early development
route,3 the introduction of this new route has shortened the
longest linear sequence from 13 steps to 9 steps and improved
the overall yield of this sequence from 4 to 16%. Overall, the
introduction of this new route has enabled the manufacture of
high-quality ceralasertib with excellent control of impurities,
resulting in processes which we believe will be suitable for long-
term commercial manufacture.
magnesium chloride (160.0 kg, 1678.4 mol) were charged to
vessel 2 at 15−25 °C. Triethylamine (192.4 kg, 1901.4 mol) was
charged to vessel 2. The contents of vessel 2 were stirred at
approximately 20 °C for 1 h. The contents of vessel 1 were
transferred to vessel 2 at 15−25 °C. The resulting mixture was
stirred at approximately 40−45 °C for 20 h. The mixture was
cooled to approximately 20 °C. Aqueous hydrochloric acid
solution (4 M, 1887 kg) was charged, and the mixture was stirred
for 1 h. Then, stirring was stopped, and the aqueous layer was
removed. Aqueous sodium bicarbonate solution (8% w/w, 799
kg) was charged, and the mixture was stirred for 30 min. Then,
agitation was stopped, and the aqueous layer was removed. A
solution of sodium chloride (184.6 kg) in water (754 kg) was
charged, and the mixture was stirred for 30 min. Then, stirring
was stopped, and the aqueous layer was removed. The organic
solution was concentrated to 2−3 relative volumes by reduced
pressure distillation. Ethanol (236.8 kg) was charged to vessel 2.
The organic solution was concentrated to 2−3 relative volumes
by reduced pressure distillation. Ethanol (236.8 kg) was charged
to vessel 2. The organic solution was concentrated to 3−4
relative volumes by reduced pressure distillation to yield ethyl 3-
[1-(methylsulfanyl)cyclopropyl]-3-oxopropanoate (20) as an
ethanol solution (552.0 kg, 37.0% w/w, 1010.5 mol, 90% yield).
1H NMR (500 MHz, DMSO, 27 °C): 1.19 (t, J = 7.1 Hz, 3H),
1.23−1.31 (m, 2H), 1.50−1.56 (m, 2H), 2.14 (s, 3H), 3.90 (s,
2H), 4.10 (q, J = 7.1 Hz, 2H).
EXPERIMENTAL SECTION
■
General Information. All starting materials were commer-
cially available and used without further purification. Large-scale
reactions were carried out in stainless steel or glass-lined steel
reactors fitted with heat transfer jackets and serviced with
appropriate ancillary equipment. NMR spectra were recorded
on a Bruker DRX 500 (500 MHz) spectrometer. The central
peaks of DMSO-d6 (δH 2.50 ppm) were used as the reference.
Sample solutions may also contain an internal standard (2,3,5,6-
tetrachloronitrobenzene) for assay determination. Spectral data
is reported as a list of chemical shifts (δ, in ppm) with a
description of each signal using standard abbreviations (s =
singlet, d = doublet, m = multiplet, t = triplet, q = quartet, br =
broad, etc.).
1-(Methylsulfanyl)cyclopropanecarboxylic acid (19).
Methyl 2,4-dibromobutyrate (351 kg, 1350 mol, 1.0 equiv) and
2-methyltetrahydrofuran (1200 kg) were charged to the vessel at
10−15 °C. A solution of sodium thiomethoxide (100.0 kg, 1427
mol, 1.06 equiv) in methanol (551.6 kg) was charged to the
vessel at 10−20 °C. The contents of the vessel were stirred at
15−25 °C for 18 h. Sodium methoxide (86.0 kg, 1592 mol, 1.18
equiv) was charged to the vessel at 10−20 °C. The contents of
the vessel were stirred at 10−20 °C for 10 h. An aqueous
solution of sodium hydroxide (2 M, 801 L, 1.22 equiv) was
charged to the vessel at 15−25 °C. The contents of the vessel
were stirred at approximately 20 °C for 12 h. The contents of the
vessel were concentrated by reduced pressure distillation to a
volume of approximately 5 relative volumes. 2-Methyltetrahy-
drofuran (1200 kg) was charged to the vessel. The mixture was
acidified to pH 1−2 with 4 molar aqueous hydrochloric acid
solution. The biphasic mixture was stirred for 1 h, and then the
batch was allowed to settle. The aqueous layer was removed.
The organic solution was washed with brine (2 × 700 kg). The
organic solution was charged with 2-methyltetrahydrofuran (2 ×
900 kg) to 3 relative volumes by reduced pressure distillation. 2-
Methyltetrahydrofuran (920 kg) was charged to the vessel. The
organic solution is concentrated under reduced pressure
distillation to yield 1-(methylsulfanyl)cyclopropanecarboxylic
acid (19) as a dry 2-methyltetrahydrofuran solution (987.2 kg,
12.0% w/w, 93% yield). 1H NMR (500 MHz, DMSO, 27 °C):
1.08−1.15 (m, 2H), 1.4−1.46 (m, 2H), 2.15 (s, 3H), 12.55 (s,
1H); 13C NMR (126 MHz, DMSO, 27 °C): 15.49, 18.98, 26.57,
173.74; HRMS (ESI): calcd for C5H7O2S [M − H]−, 131.0172;
found, 131.0159.
6-[1-(Methylsulfanyl)cyclopropyl]-2-sulfanylidene-
2,3-dihydro-4(1H)-pyrimidinone (27). Potassium tert-but-
oxide (100.0 kg, 882.4 mol, 1.2 equiv) and thiourea (67.0 kg,
880.2 mol, 1.2 equiv) charged to ethanol (710.0 kg) with stirring
at 10−40 °C was charged to the stirred solution, and the
resulting mixture was heated to approximately 78 °C. A solution
of ethyl 3-[1-(methylsulfanyl)cyclopropyl]-3-oxopropanoate
(148.0 kg, 731.7 mol) in ethanol (327.8 kg) was charged to
the vessel at 75−80 °C. The contents of the vessel were stirred at
approximately 78 °C for 23 h. The mixture was cooled to
approximately 20 °C, and then water (590 kg) was charged.
Concentrated hydrochloric acid solution (98.8 kg) was added
slowly at 15−25 °C. The contents of the vessel were stirred for 2
h. The resulting solid was collected by filtration. The filter cake
was washed with a mixture of ethanol (223.2 kg) and water
(149.0 kg) and then dried to yield 6-[1-(methylsulfanyl)-
cyclopropyl]-2-sulfanylidene-2,3-dihydro-4(1H)-pyrimidinone
(27) as a white solid (123.8 kg, 98.8% w/w, 570.7 mol, 78%
1
yield). H NMR (500 MHz, DMSO, 27 °C): 1.02−1.09 (m,
2H), 1.21−1.29 (m, 2H), 2.05 (s, 3H), 5.70 (d, J = 1.6 Hz, 1H),
12.31 (s, 1H), 12.48 (s, 1H); 13C NMR (126 MHz, DMSO, 27
°C): 14.48, 15.70, 26.54, 103.73, 155.33, 161.06, 176.31; HRMS
(ESI): calcd for C8H11N2OS2 [M + H]+, 215.0307; found,
215.0312.
6-[1-(Ethylsulfanyl)cyclopropyl]-2,4(1H,3H)-pyrimidi-
nedione (31). 6-[1-(Methylsulfanyl)cyclopropyl]-2-sulfanyli-
dene-2,3-dihydro-4(1H)-pyrimidinone (147.5 kg, 679.8 mol)
was charged to water (1066 kg) with stirring at 10−30 °C.
Chloroacetic acid (316.8 kg, 3352 mol, 4.9 equiv) was charged at
10−30 °C. The contents of the vessel were stirred at
approximately 95 °C for 6 h and then cooled to approximately
5 °C. The resulting solid was collected by filtration and then
slurried in aqueous hydrochloric acid solution (2 M, 463 kg).
The resulting solid was collected by filtration and dried at
approximately 40 °C to yield 6-[1-(methylsulfanyl)-
cyclopropyl]-2,4(1H,3H)-pyrimidinedione (31) as a white
Ethyl 3-[1-(methylsulfanyl)cyclopropyl]-3-oxopropa-
noate (20). A solution of 1-(methylsulfanyl)-
cyclopropanecarboxylic acid (148.0 kg, 1119.7 mol) in 2-
methyltetrahydrofuran (900 kg) was charged to a solution of
carbonyldiimidazole (196.0 kg, 1177.3 mol) in 2-methyltetrahy-
drofuran (764 kg) in vessel 1. The contents of vessel 1 were
stirred at approximately 20 °C for 6 h. 2-Methyltetrahydrofuran
(1615 kg) was charged to vessel 2, and then stirring was started.
Ethyl potassium malonate (285.6 kg, 1678.0 mol) and
1
solid (113.9 kg, 97.7% w/w, 561.4 mol, 83% yield). H NMR
J
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(500 MHz, DMSO, 27 °C): 1.03−1.07 (m, 2H), 1.22−1.26 (m,
2H), 2.06 (s, 3H), 5.40 (s, 1H), 10.93 (s, 2H); 13C NMR (126
MHz, DMSO, 27 °C): 14.53, 15.59, 26.69, 98.99, 151.69,
155.34, 164.05; HRMS (ESI): calcd for C8H11N2O2S [M + H]+,
199.0536; found, 199.0540.
Butyl methyl ether (651 kg) was charged to the stirred contents
of vessel 3. Aqueous NaOH solution (50% w/w, 652 kg) was
slowly charged to vessel 3 until the contents were adjusted to pH
14. Stirring was stopped, and the layers were allowed to settle.
The layers were separated, and then the aqueous layer was
recharged to vessel 3. tert-Butyl methyl ether (195 kg) was
charged to vessel 3. The mixture was stirred for 20 min, then
agitation was stopped, and the layers were allowed to settle. The
aqueous layer was removed, and then both tert-butyl methyl
ether solutions were combined in vessel 3. The contents of vessel
3 were concentrated by reduced pressure distillation to a volume
of approximately 3 relative volumes. Isopropyl alcohol (278 kg)
was charged to vessel 3, and then the contents of vessel 3 were
concentrated by distillation to a volume of approximately 3
relative volumes. Isopropyl alcohol (278 kg) was charged to
vessel 3, and then the contents of vessel 3 were concentrated by
distillation to a volume of approximately 3 relative volumes. The
mixture was cooled to 5 °C and stirred for 4 h, and then the solid
was collected by filtration and dried to yield (3R)-4-{2-chloro-6-
[1-(methylsulfanyl)cyclopropyl]-4-pyrimidinyl}-3-methylmor-
pholine (34) as a white solid (77.8 kg, 99.8% w/w, 252 mol, 68%
yield). 1H NMR (500 MHz, DMSO, 27 °C): 1.20 (d, J = 6.8 Hz,
3H), 1.23−1.31 (m, 2H), 1.49−1.58 (m, 2H), 2.11 (s, 3H), 3.19
(td, J = 13.0, 3.7 Hz, 1H), 3.44 (td, J = 11.9, 3.1 Hz, 1H), 3.59
(dd, J = 11.6, 3.1 Hz, 1H), 3.71 (d, J = 11.6 Hz, 1H), 3.82−4.09
(m, 2H), 4.33−4.37 (m, 1H), 7.10 (s, 1H); 13C NMR (126
MHz, DMSO, 27 °C): 13.37, 14.90, 20.24, 29.54, 46.73, 65.78,
69.91, 98.37, 159.69, 162.84, 170.76; HRMS (ESI): calcd for
C13H19N3SOCl [M + H]+, 300.0932; found, 300.0939.
(3R)-4-(2-Chloro-6-{1-[(R)-methylsulfinyl]-
cyclopropyl}-4-pyrimidinyl)-3-methylmorpholine (36).
Dipotassium hydrogen phosphate trihydrate (49.0 kg) and
water (1906 kg) were charged to the vessel. Stirring was started,
and then concentrated hydrochloric acid (2.2 kg) was added. A
solution of (3R)-4-{2-chloro-6-[1-(methylsulfanyl)-
cyclopropyl]-4-pyrimidinyl}-3-methylmorpholine (63.3 kg,
211 mol) and isopropyl alcohol (150.7 kg) was charged to the
vessel. Nicotinamide adenine dinucleotide phosphate (1.30 kg)
was charged to the vessel. Cyclohexanone mono-oxygenase
Rhodococcus ruber (accession number AAL14233.1, crude cell
lysate, 380.4 kg) was charged to the vessel. KRED (Asymchem
6511, 127.2 kg) was charged to the vessel. A mixture of air and
nitrogen (1:2) was blown through the reaction mixture using a
sparge, and the contents of the vessel were stirred at
approximately 30 °C for 16 h. The contents of the vessel were
adjusted to pH 3 using 10% aqueous hydrochloric acid solution
(52 kg). Sodium chloride (821.1 kg) was charged to the vessel.
The resulting mixture was stirred for 2 h. The solid was removed
by centrifugation, and the filter cake was washed with ethyl
acetate (3 × 570 kg). The combined filtrate was charged to a
clean vessel, and then a solution of sodium hydroxide (24.8 kg)
in water (101.3 kg) was added at 15−30 °C. The resulting
mixture was extracted with ethyl acetate (3 × 1170 kg). The
combined organic phases were washed with water (1232 kg).
The resulting organic solution was concentrated by reduced
pressure distillation to a total volume of approximately 250 L. n-
Heptane (434 kg) was charged to the vessel, and then the
contents of the vessel were concentrated by reduced pressure
distillation to a total volume of approximately 150 L. The
contents of the vessel were stirred at 80 °C for 2 h and then
cooled to approximately 10 °C and stirred for a further 4 h. The
solid was collected by filtration. The filter cake was washed with
heptane (2 × 170 kg) and dried to yield (3R)-4-(2-chloro-6-{1-
2,4-Dichloro-6-[1-(methylsulfanyl)cyclopropyl]-
pyrimidine (32). Phosphoryl chloride (594 kg) was charged to
vessel 1. 6-[1-(Methylsulfanyl)cyclopropyl]-2-sulfanylidene-
2,3-dihydro-4(1H)-pyrimidinone (91.7 kg, 452 mol) was
charged to the vessel with stirring at 15−20 °C. N,N-
Diethylaniline (164.8 kg, 1128 mol) was charged to the vessel
at 15−25 °C. Water (2.70 kg) was slowly charged to the vessel,
maintaining the temperature below 50 °C. The contents of the
vessel were heated at 90−100 °C for 6 h. The contents of the
vessel were cooled to 15−25 °C. A solution of sodium acetate
(18.0 kg) in water (540.8 kg) was charged to vessel 2. The
contents of vessel 1 and an aqueous solution of sodium
hydroxide (25% w/w, 2787 kg) were added to vessel 2
simultaneously, keeping the internal temperature in the range
15−30 °C and pH in the range 5−8. n-Heptane (1232.9 kg) was
charged to vessel 2, and the mixture was stirred at 35 °C for 2 h.
Stirring was stopped, and the aqueous layer was removed. A
mixture of activated carbon (9.2 kg) and water (27.2 kg) were
charged to vessel 2, and the mixture was stirred for 2 h at 35−40
°C. The mixture was filtered to remove carbon, and then the
vessel and filter cake were washed with n-heptane (124.3 kg).
The aqueous layer was removed. Water (360 kg) and aqueous
hydrochloric acid (2 M, 933 kg) were charged, the mixture was
stirred at 15−25 °C for 15 min, and then the aqueous layer was
removed. The organic solution was washed with water (2 × 270
kg) and then concentrated under reduced pressure to a volume
of 180 L. DMSO (300 kg) was charged, and then the mixture
was concentrated under reduced pressure to yield 2,4-dichloro-
6-[1-(methylsulfanyl)cyclopropyl]pyrimidine (32) as a solution
in DMSO (394.4 kg, 22.5% w/w, 377 mol, 84% yield). 1H NMR
(500 MHz, DMSO, 27 °C): 1.44−1.55 (m, 2H), 1.65−1.76 (m,
2H), 2.14 (s, 3H), 8.02 (s, 1H); 13C NMR (126 MHz, DMSO,
27 °C): 14.94, 23.02, 29.84, 117.66, 159.16, 161.92, 176.54;
HRMS (ESI): calcd for C8H9N2SCl2 [M + H]+, 234.9858;
found, 234.9863.
(3R)-4-{2-Chloro-6-[1-(methylsulfanyl)cyclopropyl]-4-
pyrimidinyl}-3-methylmorpholine (34). A solution of 2,4-
dichloro-6-[1-(methylsulfanyl)cyclopropyl]pyrimidine (87.5
kg, 372 mol) in DMSO (456.9 kg) was charged to vessel 1.
Potassium carbonate (135.0 kg, 967 mol, 2.6 equiv) was charged
to vessel 1. (R)-3-Methylmorpholine hydrochloride (62.1 kg,
447 mol, 1.2 equiv) was charged to vessel 1. Triethylamine (99.5
kg, 967 mol, 2.6 equiv) was charged to vessel 1. The contents of
vessel 1 were stirred at approximately 20 °C for 19 h. Heptane
(410.6 kg) and water (450.2 kg) were charged to vessel 1. The
contents of vessel 1 were stirred for 30 min, and then a small
amount of solid material was removed by filtration. The vessel
and the filter cake were washed with water (179 kg) and n-
heptane (278 kg), and then the layers were separated. The
organic layer was charged to vessel 2 and washed with water
(353 kg). The aqueous layer was removed. Aqueous hydro-
chloric acid solution (4 M, 1011 kg) was charged to vessel 2. The
contents of vessel 2 were agitated for 30 min, and then the layers
were allowed to settle. The aqueous layer was removed and
transferred to vessel 3. Aqueous hydrochloric acid solution (4 M,
1011 L) was charged to vessel 2. The contents of vessel 2 were
agitated for 30 min, and then the layers were allowed to settle.
The aqueous layer was removed and transferred to vessel 3. tert-
K
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
[(R)-methylsulfinyl]cyclopropyl}-4-pyrimidinyl)-3-methylmor-
pholine (36) as an off-white solid (64.2 kg, 91.8% w/w, 186.5
mol, 88% yield). 1H NMR (500 MHz, DMSO, 27 °C): 1.19 (d, J
= 6.8 Hz, 3H), 1.24−1.36 (m, 2H), 1.46 (dddd, J = 29.8, 10.5,
7.2, 5.0 Hz, 2H), 2.52 (s, 3H), 3.18 (dt, J = 12.3, 6.6 Hz, 1H),
3.42 (td, J = 11.9, 2.9 Hz, 1H), 3.56 (dd, J = 11.6, 3.0 Hz, 1H),
3.70 (d, J = 11.6 Hz, 1H), 3.92 (dd, J = 11.5, 3.5 Hz, 1H), 4.03
(q, J = 7.1 Hz, 1H), 4.28−4.32 (m, 1H), 6.66 (s, 1H); 13C NMR
(126 MHz, DMSO, 27 °C): 8.22, 12.91, 13.56, 14.05, 37.03,
42.51, 65.77, 69.87, 99.14, 158.86, 162.51, 166.08; HRMS
(ESI): calcd for C13H19N3SO2Cl [M + H]+, 316.0881; found,
316.0884.
methylsulfonimidoyl)cyclopropyl]-4-pyrimidinyl}-3-methyl-
morpholine hydrochloride (13) as a white solid (30.2 kg, 78.2
mol, 83% yield). 1H NMR (500 MHz, DMSO, 27 °C): 1.23 (d, J
= 6.8 Hz, 3H), 1.59−1.77 (m, 1H), 1.92 (d, J = 4.6 Hz, 3H),
3.16−3.22 (m, 1H), 3.42 (td, J = 11.9, 2.9 Hz, 1H), 3.57 (dd, J =
11.6, 3.0 Hz, 1H), 3.72 (d, J = 11.6 Hz, 1H), 3.78 (s, 3H), 3.94
(dd, J = 11.5, 3.6 Hz, 1H), 4.06−4.16 (m, 1H), 4.38−4.47 (m,
1H), 7.26 (s, 1H); 13C NMR (126 MHz, DMSO, 27 °C): 12.59,
12.72, 13.94, 38.57, 44.94, 65.78, 69.89, 104.70, 158.98, 160.49,
162.72; MS: HRMS (ESI): calcd for C13H20N4SO2Cl [M + H]+,
331.0990; found, 331.0991.
4-(4,4,5,5-Tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-
pyrrolo[2,3-b]pyridine (14). Isopropyl acetate (387 kg) was
charged to the vessel. The vessel was inerted with nitrogen, and
stirring was started. 4-Bromo-7-azaindole (41.5 kg, 211 mol, 1.0
equiv) was charged to the vessel. Potassium acetate (43.1 kg, 430
mol, 2.04 equiv) was charged to the vessel. Bis(pinacolato)-
diboron (54.7 kg, 215 mol, 1.02 equiv) was charged to the vessel.
Bis(triphenylphosphine)palladium(II) dichloride (2.9 kg, 4.13
mol, 0.02 equiv) was charged to the vessel. The contents of the
vessel were heated at 85−90 °C for 22 h. The mixture was cooled
to 50 °C and then washed with water (4 × 218 kg). Mercapto
silica (27.8 kg) was added to the organic phase, and the mixture
was heated at 50 °C for 8 h. The solid was removed by filtration,
and the filter cake was washed with isopropyl acetate (98 kg).
The combined filtrate was concentrated by reduced pressure
distillation to a volume of approximately 240 L. The mixture was
cooled to approximately 27 °C, and then MTBE (200 kg) was
added. The mixture was cooled to approximately 3 °C and
stirred for a further 7 h. The solid was collected by filtration. The
filter cake was washed with MTBE (40 kg) and then dried at
approximately 40 °C to yield 4-(4,4,5,5-tetramethyl-1,3,2-
dioxaborolan-2-yl)-1H-pyrrolo[2,3-b]pyridine (14) as an off-
white solid (39.7 kg, 98.9% w/w, 161 mol, 76% yield). 1H NMR
(500 MHz, DMSO, 27 °C): 1.33 (s, 12H), 6.68 (dd, J = 3.4, 1.8
Hz, 1H), 7.30 (d, J = 4.6 Hz, 1H), 7.48−7.53 (m, 1H), 8.23 (d, J
= 4.6 Hz, 1H), 11.65 (s, 1H). 13C NMR (126 MHz, DMSO, 27
°C): 24.69, 83.73, 101.37, 121.00, 123.81, 126.70, 127.56,
141.57, 147.91.
4-{4-[(3R)-3-Methyl-4-morpholinyl]-6-[1-(S-
methylsulfonimidoyl)cyclopropyl]-2-pyrimidinyl}-1H-
pyrrolo[2,3-b]pyridine (1). Anhydrous ethanol (463 kg) was
charged to the vessel and then stirring was started. (3R)-4-{2-
Chloro-6-[1-(S-methylsulfonimidoyl)cyclopropyl]-4-pyrimi-
dinyl}-3-methylmorpholine hydrochloride (41.3 kg, 111.9 mol,
1.0 equiv) and 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-
1H-pyrrolo[2,3-b]pyridine (30.1 kg, 122.5 mol, 1.1 equiv) were
charged to the vessel. The vessel was inerted with nitrogen.
Bis(triphenylphosphine)palladium(II) dichloride (1.0 kg, 1.42
mol, 0.013 equiv) was charged to the vessel. The contents of the
vessel were heated to 70 °C. A solution of anhydrous potassium
phosphate (71.0 kg, 334.5 mol, 3.0 equiv) in water (199.4 kg)
was charged to the vessel. The contents of the vessel were heated
to approximately 75 °C for 17 h. The mixture was cooled to 50
°C, and then the mixture was concentrated by reduced pressure
distillation to a total volume of approximately 5 relative volumes.
2-MeTHF (354.5 kg), water (82 kg), and L-cysteine (20.5 kg)
were charged to the vessel. The mixture was stirred at
approximately 60 °C for 4 h. Stirring was stopped, and the
aqueous layer was removed. A solution of potassium phosphate
(53.1 kg) and L-cysteine (20.8 kg) in water (124 kg) was charged
to the vessel. The mixture was stirred at approximately 60 °C for
4 h. Stirring was stopped, and the aqueous layer was removed.
(3R)-4-{2-Chloro-6-[1-(S-methylsulfonimidoyl)-
cyclopropyl]-4-pyrimidinyl}-3-methylmorpholine hy-
drochloride (13). Toluene (258 kg) and methanol (47.4 kg)
were charged to the vessel. Stirring was started, and the contents
were cooled to 0−10 °C. (3R)-4-(2-Chloro-6-{1-[(R)-
methylsulfinyl]cyclopropyl}-4-pyrimidinyl)-3-methylmorpho-
line (31.7 kg, 94.26 mol) and (diacetoxyiodo)benzene (65.0 kg,
197.9 mol) were charged to the vessel at 0−10 °C. Ammonium
carbamate (30 kg, 377.0 mol) was charged to the vessel at 0−10
°C, and then the contents of the vessel were stirred at 0−10 °C
for 20 h. Aqueous citric acid solution (30% w/w) was charged to
the vessel until the pH was adjusted to 2−3. The mixture was
stirred for 30 min. Stirring was stopped, and the aqueous layer
was removed. Stirring was restarted, and then aqueous citric acid
solution (30% w/w) was charged to the vessel until the pH was
adjusted to pH 2. Stirring was stopped, and the layers were
partitioned. The aqueous phases were combined, and stirring
was started. Aqueous sodium hydroxide solution (30% w/w)
was charged until the pH was adjusted to 8−9. Sodium chloride
(96 kg) was charged. Ethyl acetate (101 kg) and THF (33 kg)
were charged, and the resulting mixture was stirred for 30 min.
Stirring was stopped, and the layers were partitioned. The
aqueous layer was recharged to the vessel, and stirring was
started. Ethyl acetate (89.1 L) and THF (29.7 L) were charged,
and the resulting mixture was stirred for 30 min. Stirring was
stopped, and the layers were partitioned. The aqueous layer was
recharged to the vessel, and stirring was started. Ethyl acetate
(101 kg) and THF (33 kg L) were charged, and the resulting
mixture was stirred for 30 min. Stirring was stopped, and the
layers were partitioned. The aqueous layer was recharged to the
vessel, and stirring was started. Ethyl acetate (101 kg) and THF
(33 kg L) were charged, and the resulting mixture was stirred for
30 min. Stirring was stopped, and the layers were partitioned.
The organic phases were combined in the vessel and then
concentrated to approximately 59 L by distillation. Isopropyl
alcohol (48 kg) was charged, and then the solution was
concentrated to approximately 59 L by distillation. Isopropyl
alcohol (48 kg) was charged, and then the solution was
concentrated to approximately 59 L by distillation. The resulting
solution was cooled to 0−5 °C, and then a solution of hydrogen
chloride in isopropyl alcohol (6 M, 21.7 kg) was charged to the
vessel. The resulting mixture was stirred at 0−5 °C for
approximately 2 h. MTBE (135 kg) was charged to the vessel,
and the contents were stirred for a further 2 h. The solid was
collected by filtration and washed with MTBE (45 kg). The solid
was recharged to the vessel, and then methanol (54 kg) was
added. The slurry was stirred at 35−40 °C for 1 h and then
cooled to 20−25 °C. MTBE (103 kg) was charged to the vessel,
and the mixture was stirred for 1 h. The solid was collected by
filtration, and then the filter cake was washed with MTBE (59 L)
and dried to yield (3R)-4-{2-chloro-6-[1-(S-
L
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
The organic solution was washed with water (3 × 61 kg). 2-
MeTHF (534.9 kg) was charged to the vessel, and then the
solution was concentrated by reduced pressure distillation to a
volume of approximately 5 relative volumes and then cooled to 3
°C. n-Heptane (27.9 kg) was charged, and the mixture was
stirred for 2 h. The solid was collected by filtration, and then the
cake was washed with 2-MeTHF (70.7 kg) and dried at 45 °C.
The crude solid was charged to the vessel and then dissolved
in isopropanol (103.8 kg) and water (49.2 kg) at 60 °C. The
mixture was filtered to remove any solid material, and then
isopropanol (193 kg) was charged to the solution. The solution
was concentrated by reduced pressure distillation to a volume of
approximately 6 relative volumes. Isopropanol (193 kg) was
charged to the solution. The solution was concentrated by
reduced pressure distillation to a volume of approximately 6
relative volumes, until water was measured as ≤0.6%. The
mixture was held at 60 °C, and then 1 seed crystal (1.0 kg) was
charged. The stirred mixture was cooled to 25 °C and then held
at this temperature for 8 h. The stirred mixture was cooled to 5
°C and then held at this temperature for 2 h. The solid was
collected by filtration. The filter cake was washed with
isopropanol (64.7 kg) and then dried at approximately 40 °C
to yield 4-{4-[(3R)-3-methyl-4-morpholinyl]-6-[1-(S-
methylsulfonimidoyl)cyclopropyl]-2-pyrimidinyl}-1H-pyrrolo-
[2,3-b]pyridine (1) as a white solid (32.9 kg, 98.0% w/w, 78.1
mol, 70% yield). 1H NMR (500 MHz, DMSO, 27 °C): 1.27 (d, J
= 6.8 Hz, 3H), 1.4−1.47 (m, 1H), 1.47−1.6 (m, 2H), 1.76 (ddd,
J = 9.7, 7.0, 4.0 Hz, 1H), 3.11 (d, J = 1.1 Hz, 3H), 3.27 (td, J =
12.9, 3.9 Hz, 1H), 3.51 (td, J = 11.8, 3.0 Hz, 1H), 3.66 (dd, J =
11.5, 3.0 Hz, 1H), 3.79 (d, J = 11.4 Hz, 1H), 3.82−3.88 (m, 1H),
4.00 (dd, J = 11.3, 3.5 Hz, 1H), 4.19 (d, J = 10.7 Hz, 1H), 4.55−
4.61 (m, 1H), 7.00 (s, 1H), 7.23 (d, J = 2.9 Hz, 1H), 7.59 (dd, J =
3.0, 1.7 Hz, 1H), 7.96 (d, J = 5.0 Hz, 1H), 8.34 (d, J = 5.0 Hz,
1H), 11.80 (s, 1H); 13C NMR (126 MHz, DMSO, 27 °C):
11.28, 12.28, 13.45, 41.14, 46.51, 47.86, 66.04, 70.25, 101.61,
102.90, 114.65, 117.77, 127.33, 136.75, 142.29, 150.18, 161.95,
162.68, 163.28; MS (ESI) for C20H25N6O2S4 [M + H]+ 413.
Lai Chan − Chemical Development, Pharmaceutical Technology
& Development, Operations, AstraZeneca, Macclesfield SK10
2NA, U.K.
Katie G. Cooper − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Zhaoshan Cui − Asymchem Laboratories (Tianjin) Co. Ltd.,
TEDA, Tianjin 300457, P. R. China
Andrew Dalgleish − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
David Dave − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Gareth Ensor − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Rita Galan − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Peter Hamilton − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Claire Heffernan − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Lucinda V. Jackson − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Dajiang Jing − Asymchem Laboratories (Tianjin) Co. Ltd.,
TEDA, Tianjin 300457, P. R. China
Martin F. Jones − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Pengpeng Liu − Asymchem Laboratories (Tianjin) Co. Ltd.,
TEDA, Tianjin 300457, P. R. China
Keith R. Mulholland − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Mohammed Pervez − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Michael Popadynec − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Emma Randles − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.0c00482.
1
Details of H NMR spectra, 13C NMR spectra, nuclear
Overhauser effect NMR spectra, experimental details for
initial route investigation, and high-throughput exper-
imental data for the synthesis of 14 (PDF)
Simone Tomasi − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.; orcid.org/0000-0002-
9373-7639
Shenghua Wang − Asymchem Laboratories (Tianjin) Co. Ltd.,
TEDA, Tianjin 300457, P. R. China
AUTHOR INFORMATION
■
Corresponding Author
Mark A. Graham − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.; orcid.org/0000-0001-
7186-9107; Email: mark.a.graham@astrazeneca.com
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00482
Authors
Hannah Askey − Chemical Development, Pharmaceutical
Technology & Development, Operations, AstraZeneca,
Macclesfield SK10 2NA, U.K.
Andrew D. Campbell − Chemical Development,
Pharmaceutical Technology & Development, Operations,
AstraZeneca, Macclesfield SK10 2NA, U.K.
Author Contributions
The manuscript was written by M.A.G. through contributions
from all authors. All authors have given approval to the final
version of the manuscript.
M
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(11) Zenzola, M.; Doran, R.; Degennaro, L.; Luisi, R.; Bull, J. A.
Transfer of Electrophilic NH Using Convenient Sources of Ammonia:
Direct Synthesis of NH Sulfoximines from Sulfoxides. Angew. Chem.,
Int. Ed. 2016, 55, 7203−7207.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
(12) Bull, J. A.; Degennaro, L.; Luisi, R. Straightforward Strategies for
the Preparation of NH-Sulfox-imines: A Serendipitous Story. Synlett
2017, 28, 2525−2538.
■
The authors would like to thank Gair Ford, Gary Noonan, and
David McKinney for helpful discussions. We thank Simon Yates
and Holly Carter for support with high-throughput exper-
imentation and Mark Hadley and Andrew Ray for analytical
support.
(13) Grazyna, P.; Marcin, D. Alkylated derivatives of uracil. Part X.
Synthesis of 2,4-diketo-5,6,7,8-tetrahydroquinazoline. Pol. J. Chem.
1985, 59, 317−326.
(14) Hull, R.; Lovell, B. J.; Openshaw, H. T.; Payman, L. C.; Todd, A.
R. Synthetic antimalarials; some derivatives of mono- and di-
alkylpyrimidines. J. Chem. Soc. 1946, 357−362.
(15) Frutos, R. P.; Tampone, T. G.; Mulder, J. A.; Rodriguez, S.; Yee,
N. K.; Yang, B.-S.; Senanayake, C. H. Development of a Practical
Process for the Synthesis of PDE4 Inhibitors. Org. Process Res. Dev.
2016, 20, 982−988.
(16) Achmatowicz, M. M.; Thiel, O. R.; Colyer, J. T.; Hu, J.; Elipe, M.
V. S.; Tomaskevitch, J.; Tedrow, J. S.; Larsen, R. D. Hydrolysis of
Phosphoryl Trichloride (POCl3): Characterization, in Situ Detection,
and Safe Quenching of Energetic Metastable Intermediates. Org. Process
Res. Dev. 2010, 14, 1490−1500.
ABBREVIATIONS
■
API, active pharmaceutical ingredient; ATR, ataxia telangiecta-
sia mutated and Rad3 related; BVMO, Baeyer−Villiger
monooxygenase; DCM, dichloromethane; DDR, DNA damage
response; DNA, deoxyribonucleic acid; DMSO, dimethyl
sulfoxide; DIPEA, N,N-diisopropylethylamine; EtOH, ethanol;
HPLC, high performance liquid chromatography; ICH, Interna-
tional Council for Harmonization of Technical Requirements
for Pharmaceuticals for Human Use; IPA, propan-2-ol; KRED,
ketoreductase; LCMS, liquid chromatography mass spectrom-
etry; MeCN, acetonitrile; MeOH, methanol; MTBE, methyl
tert-butyl ether; MS, mass spectrometry; NADP, nicotinamide
adenine dinucleotide phosphate; NMR, nuclear magnetic
resonance; 2-MeTHF, 2-methyltetrahydrofuran; SNAr, nucleo-
philic aromatic substitution; THF, tetrahydrofuran
(17) Calculated log P for partitioning between water and n-heptane.
(18) Klamt, A. Conductor-like Screening Model for Real Solvents: A
New Approach to the Quantitative Calculation of Solvation
Phenomena. J. Phys. Chem. 1995, 99, 2224−2235.
(19) Klamt, A.; Jonas, V.; Burger, T.; Lohrenz, J. C. W. Refinement
̈
and Parametrization of COSMO-RS. J. Phys. Chem. A 1998, 102, 5074−
5085.
(20) Eckert, F.; Klamt, A. Fast solvent screening via quantum
chemistry: COSMO-RS approach. AIChE J. 2002, 48, 369−385.
(21) COSMOtherm, Version C30, Release 18; COSMOlogic GmbH
& Co. KG.
(22) Diorazio, L. J.; Hose, D. R. J.; Adlington, N. K. Toward a More
Holistic Framework for Solvent Selection. Org. Process Res. Dev. 2016,
20, 760−773.
(23) Cyclohexanone mono-oxygenase Rhodococcus ruber, accession
number AAL14233.1, as a crude cell lysate.
REFERENCES
■
(1) Foote, K. M.; Nissink, J. W. M.; Turner, P. Morpholino
pyrimidines and their use in therapy. WO 2011154737 A1, 2011.
(2) Foote, K. M.; Nissink, J. W. M.; McGuire, T.; Turner, P.;
Guichard, S.; Yates, J. W. T.; Lau, A.; Blades, K.; Heathcote, D.; Odedra,
R.; Wilkinson, G.; Wilson, Z.; Wood, C. M.; Jewsbury, P. J. Discovery
and Characterization of AZD6738, a Potent Inhibitor of Ataxia
Telangiectasia Mutated and Rad3 Related (ATR) Kinase with
Application as an Anticancer Agent. J. Med. Chem. 2018, 61, 9889−
9907.
(24) Screening details and data in Supporting Information.
(3) Goundry, W. R. F.; Dai, K.; Gonzalez, M.; Legg, D.; O’Kearney-
McMullan, A.; Morrison, J.; Stark, A.; Siedlecki, P.; Tomlin, P.; Yang, J.
Development and Scale-up of a Route to ATR Inhibitor AZD6738. Org.
Process Res. Dev. 2019, 23, 1333−1342.
(4) Mahajan, D.; Dugar, S.; Sharma, A.; Kuila, B.; Dwivedi, S.;
Tripathi, V. A Concise and Efficient Synthesis of Substituted
Morpholines. Synthesis 2014, 47, 712−720.
(5) Kanel, H. R.; Dingwall, J. D. Process for the preparation of
substituted cyclopropanecarbaldehydes. U.S. Patent 4,968,836 A, 1990.
(6) Methyl 2,4-dibromobutyrate is commercially available at 52 USD/
kg.
(7) Birch, M.; Challenger, S.; Crochard, J.-P.; Fradet, D.; Jackman, H.;
Luan, A.; Madigan, E.; McDowall, N.; Meldrum, K.; Gordon, C. M.;
Widegren, M.; Yeo, S. The Development of a Practical Multikilogram
Synthesis of the Chiral β-Amino Acid Imagabalin Hydrochloride (PD-
0332334) via Asymmetric Hydrogenation. Org. Process Res. Dev. 2011,
15, 1172−1177.
(8) Dorsey, B. D.; Dugan, B. J.; Fowler, K. M.; Hudkins, R. L.;
Mesaros, E. F.; Monck, N. J.; Morris, E. L.; Olowoye, I.; Ott, G. R.; Pave,
G. A.; Roffey, J. R. A.; Soudy, C. N.; Zificsak, C. A.; Zulli, A. L.
Azaquinazoline Inhibitors of Atypical Protein Kinase. WO 2015148597
A1, 2015.
(9) Pitchen, P.; Dunach, E.; Deshmukh, M. N.; Kagan, H. B. An
efficient asymmetric oxidation of sulfides to sulfoxides. J. Am. Chem. Soc.
1984, 106, 8188−8193.
(10) Goundry, W. R. F.; Adams, B.; Benson, H.; Demeritt, J.;
McKown, S.; Mulholland, K.; Robertson, A.; Siedlecki, P.; Tomlin, P.;
Vare, K. Development and Scale-up of a Biocatalytic Process To Form a
Chiral Sulfoxide. Org. Process Res. Dev. 2017, 21, 107−113.
N
https://dx.doi.org/10.1021/acs.oprd.0c00482
Org. Process Res. Dev. XXXX, XXX, XXX−XXX