Development and Scale-up of a Route to ATR Inhibitor AZD6738

Article abstract of DOI:10.1021/acs.oprd.9b00075

AZD6738 is currently being tested in multiple phase I/II trials for the treatment of cancer. Its structure, comprising a pyrimidine core decorated with a chiral morpholine, a cyclopropyl sulfoximine, and an azaindole, make it a challenging molecule to syn

Full text of DOI:10.1021/acs.oprd.9b00075

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC
Full Paper
The Development and Scale-up of a Route to ATR Inhibitor AZD6738
William Robert Fraser Goundry, Kuangchu Dai, Miguel Gonzalez, Daniel Legg, Anne O'Kearney-
McMullan, James Morrison, Andrew Stark, Paul S Siedlecki, Paula M. Tomlin, and Yang Jianbo
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00075 • Publication Date (Web): 21 Jun 2019
Downloaded from http://pubs.acs.org on June 22, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The Development and Scale-up of a Route to ATR
Inhibitor AZD6738
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
William R. F. Goundry,# ǂ Kuangchu Dai,‡ Miguel Gonzalez,† Daniel Legg, ǂ
Anne O’Kearney-McMullan,* James Morrison, ǂ Andrew Stark, ǂ Paul Siedlecki, *
Paula Tomlin* and Jianbo Yang†
ǂ Pharmaceutical Sciences, IMED Biotech Unit, AstraZeneca, Macclesfield, United Kingdom.
*Pharmaceutical Technology and Development, Operations, AstraZeneca, Macclesfield, United
Kingdom.
‡Changzhou SynTheAll Pharmaceutical Co. Ltd, No 589, North Yulong Road, Chunjiang Town,
Xinbei District, Changzhou 213127, Jiangsu, P. R. China
†Asymchem Laboratories (Tianjin)Co. Ltd, No. 6 Dongting, 3rd Avenue, TEDA Tianjin, 300457,
P. R. China
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 33
1
2
3
4
TABLE OF CONTENTS GRAPHIC
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
N
O
N
O
N
4 steps
80%
3 key steps
21%
Suzuki
88%
Cl
N
O
N
HN
O
N
N
HN
N
S
S
S
N
MeO
N
Cl
N
Cl
Cl
N
O
N
H
AZD6738
key steps: Biocatalytic sulfide oxidation
Sulfoximine formation
Cyclopropylation in flow
ACS Paragon Plus Environment
2

Page 3 of 33
Organic Process Research & Development
1
2
3
4
ABSTRACT
5
6
7
8
9
AZD6738 is currently being tested in multiple phase I/II trials for the treatment of cancer. Its structure,
comprising a pyrimidine core decorated with a chiral morpholine, a cyclopropyl sulfoximine and an
azaindole make it a challenging molecule to synthesize on a large scale. We describe the evolution of the
chemical processes, following the manufacture of AZD6738 from the initial scale-up through to multi-kilos
on plant scale. During this evolution, we developed a biocatalytic process to install the sulfoxide with high
enantioselectivity, followed by introduction of the cyclopropyl group firstly in batch, then in a continuous
flow plate reactor and finally through a series of continuous stirred tank reactors. The final plant scale
process to form AZD6738 was operated on 46 kg scale with an overall yield of 18%. We discuss the
impurities formed throughout the process and highlight the limitations of this route for further scale-up.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
KEYWORDS
Biocatalysis, Flow Chemistry, Suzuki, AZD6738
INTRODUCTION
The kinase ataxia telangiectasia mutated and rad3 related (ATR) is a key regulator of the DNA-
damage response (DDR) of cells, and inhibition of this repair pathway can lead to cell death.1 As
part of AstraZeneca’s investigations into blocking DDR pathways AZD6738 1, a potent and
3
selective inhibitor for ATR, was discovered (Scheme 1).2,
AZD6738 1 has excellent
physiochemical and pharmacokinetic characteristics and is currently being tested in multiple Phase
I/II trials for the treatment of both solid and hematological cancers.4 The pyrimidine core of the
molecule is decorated with a chiral methyl morpholine, an azaindole, and a sulfoximine stemming
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 33
1
2
3
4
5
6
from a cyclopropyl ring. This challenging mixture of functionality makes it an interesting target
for chemical synthesis.
7
8
9
The discovery route to AZD6738 1 started from methyl 2,4-dichloropyrimidine-6-carboxylate
2, which was reacted with (3R)-3-methylmorpholine with high selectivity for the 4-pyrimidyl
position (Scheme 1).3 The resulting pyrimidine methyl ester 3 was reduced to the alcohol 4,
activated first as the methane sulfonate 5 and then converted to iodide 6. Displacement of the
iodide by methanethiolate resulted in formation of sulfide 7. Non-stereoselective oxidation at
sulfur, followed by either crystallisation or chromatographic separation gave the desired (R)-
sulfoxide 8. Reaction with trifluoroacetamide (TFAc) catalysed by rhodium (II) gave the protected
sulfoximine 9, with subsequent installation of the cyclopropyl ring proceeding with deprotection
to give the cyclopropyl sulfoximine 10. The required boronic acid 13 was prepared from
4-bromo-7-azaindole 11 by tosyl protection to give 12, and then palladium catalyzed borylation
with bis(pinacolato)diboron. A Suzuki-Miyaura coupling of fragments 10 and 13 gave the
tosylated target 14 which was deprotected to give AZD6738 in an overall yield of 6.5%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Medicinal Chemistry Route to AZD6738 1α
ACS Paragon Plus Environment
4

Page 5 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
O
O
N
Cl
N
4 R1 = OH
5 R1 = OMs
6 R1 = I
N
N
c
d
e
N
a
b
O
N
N
Cl
R1
7 R1 = SMe
O
O
Cl
N
Cl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
3
O
O
O
N
O
N
N
f, g
N
i
h
F3C
O
O
HN
O
N
N
O
N
S
S
S
N
Cl
N
9
Cl
N
Cl
10
8
HO
OH
Br
Br
B
N
l
k
j
N
H
NTs
NTs
N
N
12
11
13
O
N
O
N
N
m
O
N
O
N
HN
HN
S
S
N
N
N
14
1
N
H
NTs
αReagent and conditions: (a) (3R)-3-methylmorpholine, TEA, DCM, 77%; (b) LiBH4, THF,
100%; (c) MsCl, TEA, DCM, 100%; (d) LiI, dioxane, 78%; (e) NaSMe, DMF, 96%; (f) m-CPBA,
DCM; (g) Crystallisation or chromatography, 40% (two steps); (h) IBDA, trifluoroacetamide,
MgO, DCM, Rh2(OAc)4 82%; (i) 1,2-dibromoethane, sodium hydroxide, TOAB, 2-MeTHF, 47%;
(j) TsCl, tetrabutylammonium hydrogen sulfate, sodium hydroxide, DCM, 92%; (k)
Bis(pinacolato)diboron, Potassium acetate, 1,1’-bis(diphenylphosphino)ferrocene dichloro
palladium(II), DMF, 62%; (l) Pd(II)Cl2(PPh3)2, Na2CO3, DME, Water, 80%; (m) 2N NaOH, DME,
water, 92%.
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 33
1
2
3
4
5
6
7
8
The medicinal chemistry route to AZD6738, while suitable for initial gram scale deliveries,
suffered from several major issues. The route was not stereoselective with ~50% of the material
being discarded at the sulfoxide stage. Both the sulfoximine formation and the cyclopropylation
were poorly understood, capricious reactions, with highly variable yields and required
chromatography. We sought to address these challenges whilst rapidly scaling up the chemistry to
meet the demands of material for clinical trials. As a secondary goal, as part of our ongoing efforts
to reduce the environmental burden of our processes, we aimed to reduce the use of solvents and
reagents particularly damaging to the environment (e.g. DCM).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Due to time pressure, the process chemistry team chose to not make a major change to the route
for Phase I/II deliveries, but to focus on finding the best conditions for each chemical
transformation. During the development of the project the team had to deliver an initial scale-up
of the medicinal chemistry route, four Kilo Lab campaigns and two pilot plant campaigns. For
clarity, the results for each synthetic stage, or group of stages where there is a reaction telescope,
are discussed in chronological order up to and including the plant manufactures.
DEVELOPMENT OF STAGES UP TO THE SULFOXIDE
For the initial scale up, (3R)-3-methylmorpholine free base was reacted with 2 under the same
conditions as the medicinal chemistry route, with removal of the solvent in vacuo giving 3 in 83%
yield (Scheme 2). The process was redeveloped for the first kilo scale manufacture to avoid the
use of DCM and to isolate 3 by a cooling crystallisation. Potassium phosphate tribasic was used
as the base in 2-MeTHF with (3R)-3-methylmorpholine again as its free base. Whilst these
conditions worked well in the lab (25 g scale, 70% yield), they did not scale-up as expected,
ACS Paragon Plus Environment
6

Page 7 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
potentially due to the thick heterogeneous reaction mixture, giving 3 in 61% (average yield for two
batches at 1.0 kg scale). The main impurity 15, observed at up to 10% area by HPLC, was formed
by substitution in the 2-position of the pyrimidine ring, with some of this carrying through to give
the reduced isomer 16 (Figure 1). From kilo scale manufacture two onwards, (3R)-3-
methylmorpholine was used as its hydrochloride salt, which allowed its addition to the reaction
mixture as an aqueous solution. We reverted to homogeneous SNAr conditions with TEA as base
in a mixture of DMSO and water. This allowed direct crystallisation of 3 from the reaction mixture
at the end of reaction. Over three manufactures, nine 3.8 kg batches were made with this process
in an average yield of 80%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthesis of sulfoxide 8β
O
N
O
N
O
N
Cl
N
N
e
a
b
O
N
N
Cl
O
S
N
R1
O
O
N
Cl
N
3
Cl
N
Cl
2
O
4 R1 = OH
5 R1 = OMs
7 R1 = SMe
8
c
d
βReagent and conditions: A First scale-up (a) (R)-3-methylmorpholine, TEA, DCM, 83%; (b)
LiBH4, THF, 96%; (c) MsCl, TEA, DCM, 100%; (d) NaSMe, water, NaI, DMF, 100%; (e) NaIO4,
water, EtOAc, MeOH then chromatography, 45%. B Kilo scale 1st Generation (a) (R)-3-
methylmorpholine, K3PO4, 2-MeTHF, 61%; (b) LiBH4, 2-MeTHF, 100%; (c) MsCl, TEA, 2-
MeTHF, 100%; (d) NaSMe, NaI, 2-MeTHF, 73%; (e) BVMO, KRED, NADP+, air, IPA, 73%. C
Kilo scale 2nd Generation (a) (R)-3-methylmorpholine HCl, TEA, Water, DMSO 80%; (b) LiBH4,
2-MeTHF, 100%; (c) MsCl, TEA, 2-MeTHF, 100%; (d) NaSMe, NaI, 2-MeTHF, 89%; (e)
BVMO, KRED, NADP+, air, IPA, 75%. D Plant scale (a) (R)-3-methylmorpholine HCl, DIPEA,
DMSO, water, 91%; (b) LiBH4, 2-MeTHF, 95%; (c) MsCl, TEA, 2-MeTHF, 100%; (d) NaSMe,
NaI, 2-MeTHF, 100%; (e) BVMO, KRED, NADP+, air, IPA, 77%.
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 33
1
2
3
4
5
6
7
8
9
O
O
Cl
N
Cl
+
N
N
–O
N
N
O
HO
N
N
N
O
S
N
N
O
O
O
O
O
S
N
17
Cl
N
Cl
15
16
18
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Significant impurities in the manufacture up to sulfoxide 8
The process to form 3 was redeveloped for the pilot plant manufacture switching the base to
DIPEA and reducing the temperature of reaction from 25°C to 0°C. This had the effect of reducing
the amount of regioisomer 15 from 8% to 4% at end of reaction. Additionally, the reaction quench
was switched from water to acetic acid which minimised degradation via ester hydrolysis. The
improvements gave a process that delivered 3 in 91% yield (average over seven 90 kg batches)
with the byproduct 15 present in 3%. The level of impurity 15 and its derivatives were tracked
through the downstream chemistry and 3% was found to be acceptably purged.
The methyl ester 3 was next quantitatively reduced to the alcohol using lithium borohydride.
The first scale-up used THF as solvent, but all subsequent kilo batches used 2-MeTHF which
allowed both an aqueous work-up and a three-step telescope to the sulfide 7. Mesylation of the
alcohol 4 was achieved by reacting 4 with methanesulfonyl chloride in the presence of TEA; the
byproduct TEA.HCl was removed by a water wash. Isolation of the intermediate iodide 6 was
avoided by using an in situ Finklestein reaction, by reacting mesylate 5 with sodium thiomethoxide
in the presence of catalytic sodium iodide. Despite extensive screening a crystalline form of 7
could not be isolated. Instead the solvent was swapped by distillation to IPA ready for the
formation of the sulfoxide.
This three-step telescope was used for ten batches of 1 – 2 kg over four separate manufacturing
campaigns with an average yield of 89%. During the kilo manufactures, upon prolonged holding,
ACS Paragon Plus Environment
8

Page 9 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
the solution of alcohol 4 was observed to become hazy indicating that crystallisation was
occurring. The opportunity was taken to improve our control strategy and attempt to purge the
regioisomer at this stage. Reducing the volume of 2-MeTHF by distillation and then adding n-
heptane as anti-solvent gave crystalline 4 in good yield. This method was used for five plant scale
batches (100 kg – 186 kg) delivering 4 in an average yield of 95% (purity 99.1% area by HPLC),
with the alcohol regioisomer 16 purged to a level of <0.2% (area by HPLC). The mesylation and
sulfide formation were then performed as a two-step telescope as before (six batches, 50 kg to 180
kg) to deliver 7 in quantitative yield.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For the first scale-up of sulfoxide 8 the oxidant was switched to sodium metaperiodate to avoid
the hazards of using m-CPBA on scale. The work up was followed by normal phase silica
chromatography and then chiral chromatography to deliver the desired sulfoxide 8 as a single
diastereomer in 45% yield. For the Kilo lab and plant manufactures we switched to enzymatic
conditions. We have previously reported in detail on the formation of the chiral sulfoxide 8 by
using a Baeyer-Villiger monooxygenase (BVMO) mediated oxidation in the presence of air.5 Key
to the successful running of the enzyme reaction on scale was modelling the mass transfer of
oxygen into the reaction mixture. We were then able to combine a high enough stirring rate with
the sparging of air to ensure that the enzyme did not become oxygen starved. Both the over
oxidised product, sulfone 17, present at 0.3% (area HPLC), and the N-oxide 18, found in 3.4%
(area HPLC) were successfully removed by crystallisation from 2-butanone (Figure 1). Following
our publication, the enzyme system has been used to manufacture six more plant scale batches (up
to 95 kg) in an average yield of 79% and a chiral purity of >99.7% (area by HPLC).
DEVELOPMENT OF THE SULFOXIMINE FORMATION AND CYCLOPROPANATION
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 33
1
2
3
4
5
6
7
8
9
For the initial scale-up to form the sulfoximine functionality, the conditions reported by the
group of Bolm utilising a rhodium catalyzed nitrene formation and insertion were followed.6 The
reaction used trifluoroacetamide (TFAc) as the electron deficient nitrogen source giving, after
chromatography, a 93% yield of trifluoroacetamide protected sulfoximine 9 as an orange gum
(Scheme 3). This was then reacted with 1,2-dibromoethane under phase transfer conditions to form
the cyclopropyl ring with concomitant removal of the TFAc protecting group. This reaction was
slow to proceed (three days), gradually going black over time. Following chromatography, a 60%
yield of cyclopropyl sulfoximine 10 was isolated, representing a 56% yield from sulfoxide 8.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3. Formation of the sulfoximine and cyclopropyl ringγ
O
O
N
O
N
b
N
a
F3C
O
O
N
HN
O
N
N
O
S
N
S
S
N
Cl
N
9
Cl
N
Cl
10
8
O
N
c
d
O
N
HN
S
N
Cl
19
γReagent and conditions: A First scale-up (a) IBDA, trifluoroacetamide, MgO, Rh2(OAc)4,
DCM, chromatography, 93%; (b) batch, NaOH, 1,2-dibromoethane, TOAB, 2-MeTHF,
chromatography, 60% (c) NA (d) NA; B Kilo scale 1st generation (a) IBDA, trifluoroacetamide,
MgO, Rh2(OAc)4, Anisole, telescope; (b) NA; (c) H2SO4, iPrOAc, chromatography, 40%; (d) flow,
KOH, 1-bromo-2-chloroethane, TBAB, anisole, chromatography, 74%; C Kilo scale 2nd
generation (a) IBDA, trifluoroacetamide, RhAc dimer, Anisole, telescope; (b) NA; (c) KOH,
anisole, 78%; (d) flow, KOH, 1-bromo-2-chloroethane, TBAB, anisole, chromatography 70%; D
Kilo scale 3rd generation (a) IBDA trifluoroacetamide, Rh2(OAc)4, anisole, 85%; (b) KOH,
ACS Paragon Plus Environment
10

Page 11 of 33
Organic Process Research & Development
1
2
3
4
5
6
1-bromo-2-chloroethane, TBAB, anisole, chromatography, 55%; (c) NA (d) NA; E Plant scale (a)
IBDA, trifluoroacetamide, Rh2(OAc)4, anisole, 79%; (b) NA; (c) Conc HCl, anisole, 92%; (d)
CSTR; KOH, 1-bromo-2-chloroethane, TBAB, 2-MeTHF then IPA, HCl, TBME, 41% as HCl
salt.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For the first Kilo scale manufacture we decided to investigate the formation and isolation of the
free sulfoximine 19. Based on the hypothesis that the yield for the chemistry to install the
cyclopropyl was linked to the nature of the protected sulfoximine intermediate, we sought to
optimize the Bolm conditions and combine them with a deprotection step in the reaction work-
up.7-11 Performing the reaction in anisole instead of DCM allowed us to heat the reaction to 80°C,
reducing the reaction time to two hours. Once the reaction was cooled, solid material could be
removed by filtration. Aqueous 2M sulfuric acid was then used to deprotect the trifluoroacetamide
group, and the organic layer was removed. The aqueous could then be neutralised, the product
extracted into isopropyl acetate and crystallised. Whilst this approach worked well on a lab scale
(~70% yield), upon scaling to 1 kg the yield reduced to 40%. It was proposed that the yield
reduction was caused by the degradation of 19 due to prolonged hold in the acidic deprotection
solution.
For the cyclopropanation reaction, previous data from the first scale-up under batch conditions
implied that the product was slowly decomposing over time. To hopefully avoid this, flow
conditions were developed for the first kilo batch to allow us to perform the reaction at an elevated
temperature for a limited time. There were several key challenges to overcome to develop the
phase transfer reaction in flow: understanding the mixing of the aqueous and organic streams;
potential formation of inorganic solids; potential formation of vinyl halides as gaseous byproducts.
Initial parallel experimentation in batch identified lead reaction conditions of sulfoximine 19 in
anisole, with aqueous potassium hydroxide, TBAB and 1-bromo-2-chloroethane. A series of
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 33
1
2
3
4
5
6
7
8
9
Design of Experiments (DOE) were then carried out on an Uniqsis Flowsyn reactor, fitted with an
IMM static mixer, investigating: the ratios of aqueous to organic streams; reaction temperatures;
residence times and types of emulsions formed (temporary or semi-permanent). The work on the
Flowsyn did not give full understanding of the process but did give us confidence the process
would work on scale. The IMM static mixer was also frequently blocked by solid potassium
bromide formation. For the first Kilo manufacture aqueous potassium hydroxide was pumped
through an Alfa Laval flow reactor unit and the organic stream was sprayed into it through a nozzle
with only minor solid formation (Figure 2). The reaction progressed to give, under steady state
conditions, 80% cyclopropyl 10 and 10% overreaction impurity 20 (Figure 3). Following work-up
the desired compound 10 was separated by chromatography and isolated in 74% yield (0.66 kg),
representing a 30% yield from sulfoxide 8.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Schematic for the Alfa Laval flow manufacture of 8 at kilo scale. Reactor set-up: A Kilo
scale 1st generation: seven reactor plates heated to 60°C, anisole:KOH flow rate ratio varied
between 1:1.3 and 1:1.5, residence time of 2.8 to 3.0 min with a pressure drop of 2.6 bar.12 B Kilo
scale 2nd generation: seven reactor plates heated to 60°C, anisole:KOH flow rate ratio 1:1.8,
residence time of 3.0 min with a pressure drop of ~2.0 bar.13
ACS Paragon Plus Environment
12

Page 13 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
O
N
O
N
Cl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
N
O
N
N
HN
S
S
N
Cl
N
O
20
23
Figure 3. Significant impurities in the formation of 10
For the second Kilo scale manufacture, to avoid yield loss during the removal of the
trifluoroacetamide group to afford 19, an alternative nosyl protecting group was envisioned (21,
Figure 4). It was proposed that the nosyl group could be carried through several stages of
chemistry, perhaps conferring improved crystallinity, before deprotection later in the endgame.
The required intermediate iminoiodinane 22 was prepared following the literature method.14
However, safety testing showed the compound to be highly energetic in nature and potentially
explosive, so this route was not explored further for use at scale.
Work then focused on improving the existing conditions for the formation of 9. The use of iron
catalysts as described in the literature was investigated using both iron(III)tris(acetylacetonoate)
and iron(III) trifluoromethane sulfonate.15, 16 No conversion to the desired products was observed
coupled with a large amount of degradation. Surprisingly, during a base screen to find an
alternative to magnesium oxide, the control reaction with no base led to 9 in high yield. The TFAc
group could then be removed by stirring with 4M KOH at 20˚C and the anisole solution used
directly in the flow chemistry. Two batches of sulfoximine 19 were manufactured using these
conditions (78% average yield), although some degradation was observed.
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 33
1
2
3
4
5
6
7
8
O
N
O2N
O2N
9
O
N
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
S N
S
|
O
O
S
O
O
N
Cl
22
21
Figure 4. The proposed Nosyl sulfoximine 21 and Nosyl iminoiodinane 22
The degradation caused by the KOH deprotection had a negative impact on the flow formation
of the cyclopropane 10, including the precipitation of potassium bromide in the flow reactor. To
mitigate this issue the reaction path was shortened from nine plates to seven plates, reducing the
residence time from four minutes to three minutes. Two batches were manufactured using this
process giving cyclopropane 10, after work-up and chromatography, in 70% yield, representing a
55% yield from sulfoxide 8. The second batch, using telescoped sulfoximine 19, proved very
capricious with numerous blockages and issues due to solid build up in the reactor.
For the final Kilo manufacture we investigated forming the sulfoximine 19 directly from 8 with
no protecting group using hydrazoic acid generated in situ in flow.17 Whilst the reaction worked
well to give 77% 19, the chirality at the sulfur was not retained, hence the process was not used
for manufacture. Instead, to avoid the degradation issues resulting from the KOH deprotection
step, the TFAc group was not removed, but a solution of TFAc sulfoximine 9 in anisole was used
directly in the cyclopropanation. The manufacture reverted to batch conditions, with both removal
of the TFAc group and formation of the cyclopropyl occurring in tandem. The material was
purified by chromatography and then recrystallized from n-butanol to give cyclopropyl 10 in 47%
yield over two steps.
ACS Paragon Plus Environment
14

Page 15 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
On plant scale the reaction conditions to form the sulfoximine remained unchanged, but acidic
work-up conditions were used to remove the TFAc group, as sulfoximine 19 was more stable at
an acidic pH than basic and could be extracted into 2-MeTHF ready for the next stage. Six batches
were manufactured on a 76 kg scale with an average yield of 71%. The cyclopropanation was then
achieved by using a continuous stirred tank reactor (CSTR) with four reactors in series followed
by a batch work-up (Figure 5). The residence time of the CSTR was balanced to ensure high
product conversion while minimizing the formation of the over alkylated impurity 20 to <10%.
Following aqueous work-up, the solvent was switched by distillation to isopropanol and 10 was
isolated as the HCl salt, with impurity 20 at ~1.4% (area by HPLC). A lab trial recrystallisation of
the HCl salt in methanol reduced the level of 20 to ≤0.4% (area by HPLC) but introduced the new
impurity 23 by methanol displacement of chloride (Figure 3). However, keeping the temperature
below 55°C the purification could be achieved avoiding formation of 23. While this process on
plant scale produced material of high purity (>99% area by HPLC), the combined yield for the
CSTR and recrystallisation was relatively low compared to the Kilo lab (8 batches, average yield
41%), representing a 27% yield from sulfoxide 9.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
15

Organic Process Research & Development
Page 16 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Schematic for the CSTR manufacture at plant scale. Four stage CSTR cascade, residence
time up to 2.0 h. Batch work-up to give 10 as the HCl salt.
DEVELOPMENT OF FINAL STAGES TO AZD6738
During the initial scale-up of the boronate ester fragment, it was recognised that the tosyl
protecting group could be avoided (Scheme 4). Deprotonation of the commercially available
4-bromo-7-azaindole 11 with sodium hydride, followed by bromide lithium exchange at -74°C
with n-BuLi and quenching with trisopropyl borate, gave a mixture of boronic acid 24 and boronate
ester. This mixture could be isolated by direct solvent removal and used in the final coupling
without purification. However, as the material was a final isolated intermediate in the synthesis,
we preferred a defined crystalline material rather than a mixture of species. For the kilo scale
manufactures we switched the borylating agent to isopropyl pinacol borate which allowed the
crystallisation of boronate ester 25 from n-butanol. While the yield was reduced, the material was
ACS Paragon Plus Environment
16

Page 17 of 33
Organic Process Research & Development
1
2
3
4
5
6
a crystalline single compound and this process was used to complete 13 batches on a 1.6 kg scale
(average yield 53%).
7
8
9
Scheme 4. Formation of the azaindole coupling partner δ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R1
R2
Br
B
N
N
H
N
N
H
11
24 R1=R2= OH
25 R1=R2= Pinacol
δReagent and conditions: A First scale-up, NaH, THF, 5°C, then n-BuLi, -74°C, triisopropyl
borate, 94% 24; B Kilo scale, NaH, THF, 5°C, then n-BuLi, -74°C, isopropyl pinacol borate, 53%
25; C Plant scale, bis(pinacolato)diboron, Pd(PPh3)4, potassium acetate, 1,4-dioxane, 90°C, 63%,
25.
For plant scale, to avoid the low temperatures required for the lithiation chemistry, the process
was switched to a palladium catalyzed borylation using tetrakis(triphenylphosphine)palladium(0)
with bis(pinacolato)diboron 1,4-dioxane giving boronate 25 in good yield (Scheme 4).18 Efficient
purging of oxygen from the vessel was key to avoid dimerization byproducts and keeping the
homocoupled impurity 26 to below 0.8% (Figure 6). A telescope of the borylation chemistry
into the final Suzuki coupling was investigated. Upon completion of the borylation, by adding
the aryl chloride 10 and potassium carbonate, AZD6738 1 could be accessed directly in good
yield. However, this would have placed a high purification burden on the final workup and
isolation and so the telescope was not used for manufacture. The palladium catalysed borylation
was used to manufacture four plant batches of 25 (30 – 60 kg) in an average yield of 63%.
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 33
1
2
3
4
5
6
7
8
9
As the tosyl protection was no longer required, the final stage of the synthesis became the Suzuki
coupling. The initial scale-up used the same conditions as the medicinal chemistry route, giving
after chromatography, AZD6738 1 in 70% yield. To develop more process friendly conditions and
reduce the potential for palladium carryover, a process was developed using the 3-(Di-t-
butylphosphonium)propane sulfonate (DTBPPS) palladium system with n-butanol as solvent.19
DTBPPS is a water-soluble ligand, and so the palladium can largely be removed from the system
via aqueous washes. The DTBPPS catalyst system was very sensitive to oxidation, and the first
two lab reactions stalled (around 50%) due to incomplete inertion, however when rigorous inertion
was used, all reactions proceeded in >95% (area by HPLC) conversion. To ensure robustness, the
level of catalyst was increased from 0.3 mol% to 0.5 mol%, however during the first Kilo scale
manufacture (single batch strategy) the reaction stalled at 5% conversion and could not be restarted
by the adding more catalyst or base. Fortuitously, intermediate 10 could be recovered from the
batch and sufficient boronate ester 25 was available to repeat the reaction. The team then reverted
to the Pd(PPh3)2Cl2 conditions in n-butanol, which allowed good conversion to be achieved
(Scheme 5). Post work-up the palladium level was reduced by use of solid supported scavenger
and AZD6738 1 was crystallised in 50% yield; however, 25% of the material remained in the
mother liquors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Formation of AZD6738 ε
ACS Paragon Plus Environment
18

Page 19 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
O
N
O
N
O
O
O
N
HN
a
B
N
S
N
+
O
N
HN
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
S
N
Cl
N
H
N
H
10
25
1
εReagent and conditions: A First scale-up (a) 24 1.30 equiv, Pd(PPh3)2Cl2 (5 mol%), NaHCO3,
DME, water, 90°C, 70%; B Kilo Scale (a) 25 1.30 equiv, Pd(PPh3)2Cl2 (5 mol%), K2CO3, n-
Butanol, 90°C, 52-83%; C Plant scale (a) Starting material 10, as HCl salt; 25 1.1 equiv,
Pd(dppf)Cl2 (2 mol%), K2CO3, ethanol, 75°C, 65-100%.
For the next two Kilo campaigns the process remained the same with only minor modifications
to the crystallisation (isohexane as anti-solvent) which increased the yield to 74% (average of 4
~1.8 kg batches). However, for these batches the level of the homo-coupled impurity 26 was close
to the specification limit of ≤1.3% (area by HPLC), and this impurity was difficult to purge in the
work-up and crystallisation (Figure 6). For the final Kilo scale manufacture, we focused on
developing the process to reduce the level of this impurity. The work of Lloyd-Jones and co-
workers has shown that the level of homo-coupled dimer is linked to both the level of oxygen
present and the catalyst loading,20 with each mole of oxygen present affording 1 mole of 26 and
each mole of Pd(II) catalyst furnishing an additional 1 mole of 26. However, when the level of
Pd(PPh3)2Cl2 was reduced from 5 mol% to 1 mol% the reaction was significantly slower.
Maintaining the catalyst loading at 2 mol%, the reaction temperature could be reduced from 90°C
to 75°C, with the level of the homo-coupled impurity 26 dropping to 0.04% in the isolated product.
The final kilo manufacture using the new conditions produced two 2.7 kg batches in an average
yield of 80%.
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 33
1
2
3
4
5
6
7
8
9
O
N
H
N
N
Cl
O
N
N
S
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
N
N
H
N
H
26
27
Figure 6 Significant impurities in the formation of AZD6738 1
Prior to plant scale manufacture, a catalyst and ligand screen identified Pd(dppf)Cl2 as a superior
catalyst. The reaction used the crystalline HCl salt of 10 as the input material necessitating the use
of an additional equivalent of potassium carbonate. The reaction could be run with only 1.1
equivalents of boronate ester 25 and in ethanol as solvent, giving a clean reaction profile and
complete reaction in two hours. The work-up and crystallisation remained unchanged, with the
solvent swapped to n-butanol following the catalyst removal by filtration. The input material 10
for the first plant batch had an elevated level of impurity 20, which reacted to give the related
impurity 27 above the specification level. The batch was recrystallized to remove 27, prioritizing
quality over yield, giving only 7.1 kg, in 41% yield. Subsequent plant batches, with lower levels
of impurity 20 in the input 10, proceeded as expected giving AZD6738 1 in an average yield of
88% yield (27.8 kg, 39.8 kg and 46.2 kg).
A rework process was developed to reduce the level of residual palladium if required. AZD6738
1 could be dissolved in n-butanol, treated with a silica thiol based scavenger then reisolated by
crystallisation in an essentially quantitative yield. This process was used in the final two plant
batches and was successful in reducing the level of palladium from ~180 ppm down to 15 ppm.
ACS Paragon Plus Environment
20

Page 21 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The history of our development is summarized in Table 1. This shows that our ‘improvements’
did not always lead to immediate success but demonstrated that the scale-up of chemical processes
is an iterative process, building on the learning of previous manufacturing campaigns. This was
especially true with this molecule for the introduction of the challenging chiral sulfoximine and
the cyclopropyl groups. The reduction in yield for the first kilo manufacture was attributable to
both the diminished yield of the sulfoximine formation and high losses to the liquors in the
crystallisation of AZD6738. The reduction in yield on moving to the plant reflected the complexity
of tech transferring the biocatalytic process to this scale, and the continuing development of the
CSTR process.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. A summary of the scale-up manufactures of AZD6738 θ
Scale
Initial
Kilo 1 Kilo 2 Kilo 3 Kilo 4 Plant 1 Plant 2
scale-up
Overall
(%)ι
Yield 14
4
20
18
25
9
18
Largest batch of 0.28
AZD6738 (kg)
0.31
2.00
2.41
2.75
27.8
46.2
θKilo 3 was a repeat of Kilo 2 and no development work was performed between these deliveries
ιYield for the longest sequence starting from 2 to 1
CONCLUSION
We scaled the production of AZD6738 to a multi-kilo plant manufacture incorporating a
biocatalytic stage and a cyclopropanation in continuous stirred tank reactors; overall, we have
produced 12 batches at an increasing scale over time. Key improvements were the development of
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 33
1
2
3
4
5
6
crystalline isolated intermediates, minimising the use of harmful solvents (DME, DCM), and
reducing the use of chromatography.
7
8
9
The manufacture of AZD6738 remains a challenge for the future of this medicine, due to the
difficult nature of installing the dense functionality around the pyrimidine core. Longer term it
would be beneficial to have a more convergent approach with a higher yielding route and our
efforts in this area have begun.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL
Synthesis of methyl 2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidine-4-carboxylate 3;
methyl 2,6-dichloropyrimidine-4-carboxylate 2 (86.5 kg, 418 mol, 1.00 equiv) and (3R)-3-
methylmorpholine hydrochloride (63.3 kg, 460 mol, 1.10 equiv) were charged to a stirred solution
of DMSO (923 kg, 10.6 rv) and water (174 kg, 2 rv). The reaction was cooled to 0-5°C, N,N-
diisopropylethylamine (114 kg, 1127 mol, 2.70 equiv) and water (174 kg, 2 rv) were charged, then
the reaction was held at 0-5°C for 5 h. Acetic acid (90 kg, 5% aq) was added to neutralise the
reaction, maintaining the temperature at 0-10°C. The reaction was then heated to 60-65°C, and
water (519 kg, 6 rv) was added precipitating the product. The slurry was cooled back to 20-25°C
and the solid isolated via centrifuge and washed with water (432 kg, 5 rv). The solid was dried to
a constant weight at 50-55°C to give methyl 2-chloro-6-[(3R)-3-methylmorpholin-4-
yl]pyrimidine-4-carboxylate 3 as an off-white solid (96.5 kg, 97.9%w/w, 83% yield); Karl Fisher
0.1%; Assay by HPLC 96.7%; Chiral purity 99.3%.
Synthesis of [2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-yl]methanol 4; methyl 2-
chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidine-4-carboxylate 3 (186 kg, 685 mol 1.00 equiv)
ACS Paragon Plus Environment
22

Page 23 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
was dissolved in 2-MeTHF (980 kg, 6.0 rv) and cooled to 0-5°C with stirring. To this was added
a solution of lithium borohydride (197.5 kg, 2M in THF, 0.65 equiv) and the reaction was held at
0-5°C for 11 h. Acetic acid (257 kg, 10% aq, 0.65 equiv) was added followed by potassium
bicarbonate (324 kg, 20% aq, 1.00 equiv). The bottom aqueous layer was separated and discarded
then the organic layer was washed with sodium sulfate (398 kg, 7% aq). The volume of the organic
layer was reduced by distillation at 40°C to ~2-3 rv, and the temperature was cooled to 20-25°C.
n-heptane (334 kg, 1.8 rv) was added precipitating the product which was then isolated via
centrifuge and washed twice with n-heptane (2 × 74 kg, 0.5 rv). The solid was dried to a constant
weight to give [2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-yl]methanol 4 as an off-
white solid (158.3 kg, 95.1% yield); Karl Fischer 0.3%; Assay by HPLC 100.2%; Chiral purity;
99.9%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of (3R)-4-[2-chloro-6-(methylsulfanylmethyl)pyrimidin-4-yl]-3-methyl-morpholine
7; To a stirred solution of [2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-yl]methanol 4
(156.3 kg, 641 mol, 1 equiv) in 2-MeTHF (782.2 kg, 7.3 rv) at 0-5°C was added triethylamine
(90.4 kg, 893 mol, 1.35 equiv). Methanesulfonyl chloride (90.4 kg, 789 mol, 1.20 equiv) was
added and the reaction was held at 0-5°C for 3 h. The organic solution was washed twice with
potassium phosphate monobasic aqueous (2×102 kg, 2.5% aq), keeping the temperature at 0-5°C,
then once with sodium sulfate (52.0 kg, 7% aq) at 5-15°C. The organic layer was cooled to 0-5°C,
sodium thiomethoxide (296 kg as a 20%wt/wt solution, 1.30 equiv) was added and the reaction
was held at 0-5°C for 3 h. The reaction was washed three times with sodium sulfate (2×148 kg,
7% aq) and then the organic layer was warmed to 15-25°C. The organic layer was washed with
potassium phosphate monobasic (52 kg, 5% aq), then the solvent was switched from 2-MeTHF to
ACS Paragon Plus Environment
23

Organic Process Research & Development
Page 24 of 33
1
2
3
4
isopropanol by distillation and the addition of isopropanol (259 kg, followed by 161 kg). This gave
5
6
(3R)-4-[2-chloro-6-(methylsulfanylmethyl)pyrimidin-4-yl]-3-methyl-morpholine
7
as an
7
8
9
isopropanol solution (797 kg, 100% yield); Karl Fischer 0.24%; Assay by HPLC 25%; Chiral
purity; 99.9%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis
of
(3R)-4-[2-chloro-6-[[(R)-methylsulfinyl]methyl]pyrimidin-4-yl]-3-methyl-
morpholine 8; A solution of (3R)-4-[2-chloro-6-(methylsulfanylmethyl)pyrimidin-4-yl]-3-methyl-
morpholine 7 (92.4 kg, 264 mol., 1.00 equiv) in isopropanol (274 kg, 3.0 rv) was added to a mixture
of water (4547 kg), triethanolamine (34.2 kg, 0.37 equiv) and 1N HCl (18.5 kg, 0.20 equiv) in
reactor one. Following a 2-propanol line wash (15.6 kg) the reaction was heated to 24-27°C.
β-triphosphopyridine (1.0 Kg, 0.01 equiv), KRED CDX-019 Enzyme (2.0 kg. 2.04 equiv) and
BVMO enzyme (4.2 kg, 0.02 equiv) were charged to reactor one. Water (585 kg) and 2-propanol
(26 kg) were charged to reactor two. Compressed air was bubbled through reactor two into reactor
one at a rate of 2-4 m3/h. The reaction was stirred at 24-27°C for 60 h during which time a further
100 kg of water was added to reactor two. Following reaction completion 2-butanone (289 kg, 4.1
rv) was charged to reactor one and the reactor volume was reduced to ~1400 L by distillation under
reduced pressure below 45°C. 2-butanone (827 kg) was charged to reactor one, the temperature
adjusted to 48-50°C, and the reaction stirred for 30 minutes. Following settling the aqueous layer
was removed and retained. The organic layer was transferred to reactor three and reduced in
volume to ~120 L by distillation under reduced pressure, below 45°C. The 2-butanone distillates
and aqueous layer were recharged to reactor two and the extraction and distillation repeated; this
process was repeated four times. To the resulting 2-butanone reaction mixture in reactor three was
charged DCM (1884 kg) and the reaction was stirred for 30min. The organic layer was separated
and retained. The aqueous layer was washed with further DCM (470 kg) and the aqueous layer
ACS Paragon Plus Environment
24

Page 25 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
discarded. 2-butanone (267 kg) was charged to the combined organic layers in reactor three and
the volume reduced to 124 L by distillation under reduced pressure, below 45˚C. Further 2-
butanone (514 kg) was charged and the volume reduced to ~250 L by distillation under reduced
pressure, below 45°C. 2-butanone (267 kg) was charged to reactor three and the reaction was held
at ~60°C for 1 h. The reaction mixture was cooled over 10.5 h to ~-10˚C and held for 5 h. The
crystalline product was isolated by centrifuge and washed twice with pre-cooled 2-butanone (2×57
kg, ~-10˚C). The wet cake was dried for 36 h at 45°C under reduced pressure and discharged to
give (3R)-4-[2-chloro-6-[[(R)-methylsulfinyl]methyl]pyrimidin-4-yl]-3-methyl-morpholine 8 as
an off white solid (62.1kg, 70% yield); Assay (HPLC) 99%; Chiral purity (HPLC) (R,R)-
diastereoisomer 99.7%, (R,S)-diastereoisomer 0.3%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of N-[[2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-yl]methyl-methyl-oxo-
λ6-sulfanylidene]-2,2,2-trifluoro-acetamide 9; to a stirred solution of (3R)-4-[2-chloro-6-[[(R)-
methylsulfinyl]methyl]pyrimidin-4-yl]-3-methyl-morpholine 8 (77.6 kg, 1.00 equiv), in anisole
(736 kg) was charged Iodobenzene diacetate (138.0 kg, 1.60 equiv), 2,2,2-trifluoroacetamide (41.1
kg, 1.35 equiv) and a solution of rhodium(II)acetate dimer (1.30 kg, 0.11 equiv) in anisole (5.0 L).
The mixture was heated to ~60°C and held for 3.3 h. The reaction mixture was cooled to ~25°C,
filtered and then washed with water (2 × 237.3 kg) twice, to give N-[[2-chloro-6-[(3R)-3-
methylmorpholin-4-yl]pyrimidin-4-yl]methyl-methyl-oxo-λ6-sulfanylidene]-2,2,2-trifluoro-
acetamide 9 as an anisole solution (88.3 kg, 83.3% yield); Assay (HPLC) 85%.
Synthesis of [2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-yl]methyl-imino-methyl-
oxo-λ6-sulfane 19; to a stirred solution of N-[[2-chloro-6-[(3R)-3-methylmorpholin-4-
yl]pyrimidin-4-yl]methyl-methyl-oxo-λ6-sulfanylidene]-2,2,2-trifluoro-acetamide 9 (172.7 kg,
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 33
1
2
3
4
5
6
7
8
9
1.00 equiv) in anisole (1855 kg) at ~10°C was added HCl (203.2 kg, conc. aq). After 1 h the
aqueous layer was removed and retained and a further HCl (203.2 kg, conc. aq) was charged and
again stirred for 1 h. The organic layer was removed and discarded and the combined aqueous
layers were adjusted to pH ~8 by the addition of potassium carbonate solution (496.2 kg in 647.7
kg water). The reaction solution was heated to ~35°C and extracted three times with 2-MeTHF (3
× 217.0 kg). The combined organic extracts were washed with sodium chloride solution (43.0 kg
in 172.1 kg water) to give [2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-yl]methyl-
imino-methyl-oxo-λ6-sulfane 19 as a dark red 2-MeTHF solution (124.7 kg, 85% yield); Assay
(HPLC) 89.8%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of [1-[2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-yl]cyclopropyl]-imino-
methyl-oxo-λ6-sulfane 10 hydrochloride salt; To a stirred solution of [2-chloro-6-[(3R)-3-
methylmorpholin-4-yl]pyrimidin-4-yl]methyl-imino-methyl-oxo-λ6-sulfane 19 (56.6 kg, 1.0
equiv) in 2-MeTHF (581.8 kg) was added tetrabutylammonium bromide (14.9 kg, 0.25 equiv) and
1-bromo-2-chloroethane (120.1 kg, 4.5 equiv). The mixture was stirred until all solids had
completely dissolved to form the organic feed solution. A 50% KOH aqueous solution (571.3 kg
with 570.0 kg water) was prepared as feed solution B. Four CSTRs in series were used for this
reaction. Each reactor had a total volume of 20 L but were operated at 14 L to achieve a residence
time between 1.5 to 2 h. The CSTRs were equipped with Pitched Blade Turbine (PBT) impellers
and baffles to enhance reagents’ mixing. Transfer of material from one CSTR to the next was by
gravity (i.e. overflow). Each reactor jacket was set at ~70°C. Feed solutions A and B were pumped
in the first of the CSTR at a rate to ensure a total residence time of 1.5-2 h. After the last reactor
the reaction mixture flowed into a receiving vessel at 15°C. Water (280.0 kg) was charged to
dissolve solid material, then the aqueous layer was removed and retained. The organic solution
ACS Paragon Plus Environment
26

Page 27 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
was washed four times with citric acid (4 × 100 kg, 5% aq) with the pH being adjusted to ~pH 1.0
by the addition of HCl (1M); each aqueous wash was back extracted with ethyl acetate (608.7 kg),
and all the organic layers were discarded. The combined aqueous layers were neutralized with
potassium hydroxide (50% aq) maintaining the temperature at ~20°C. The product was then
extracted by washing with ethyl acetate (4 × 203.0 kg), and then concentrated to ~3 rel vol to give
10 as a dark brown solution in ethyl acetate. The solution was azeotroped twice with isopropanol
(put and take, 120.4 kg then 386 kg) to reduce the water level to below 2% (as measured by Karl
Fischer). The mixture was cooled to 20°C and HCl (38.6 kg as a 15% solution in IPA) was added
at the rate of ~20 kg/h. The reaction mixture was held for 1 h, then MTBE (169.4 kg) was charged
at the rate of 60 kg/h, then the mixture was cooled to ~5°C and held for 1 h. The solid was filtered
and rinsed with MTBE (56.4 kg) and pulled dry under vacuum. The solid was redissolved in
methanol (180.8 kg) at 50°C then cooled back to 20°C. MTBE (61.4 kg) was charged and the
mixture was cooled to ~5°C and held for 1h. The solid was collected by filtration, rinsed with
MTBE and dried to give [1-[2-chloro-6-[(3R)-3-methylmorpholin-4-yl]pyrimidin-4-
yl]cyclopropyl]-imino-methyl-oxo-λ6-sulfane 10 hydrochloride salt (31.4 kg, 51.1%) as a white
solid; HPLC Purity 99.0%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-pyrrolo[2,3-b]pyridine 25;
1,4-dioxane (900.6 kg) was charged to a reactor and degassed with nitrogen for 2 h. To the stirring
solution at ~25°C were added, 4-bromo-7-azaindole 11 (60.0 kg, 1.0 equiv), potassium acetate
(60.3 kg, 2.0 equiv), bis(pinacolato)diboron (100.9 kg, 1.3 equiv), and
tetrakis(triphenylphosphine)palladium(0) (3.5 kg, 0.01 equiv). The mixture was heated to ~95°C
and held at this temperature for 12 h. The mixture was cooled to 30°C and solids removed by
filtration, rinsing the solids with 1,4-dioxane (100.4 kg). The organic filtrates were combined and
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 33
1
2
3
4
5
6
7
8
9
concentrated under reduced pressure, keeping the temperature below 60°C, down to 2.0 rel vol.
The mixture was heated to 35°C and then water (180.5 kg) was added followed by cooling to 20°C
and holding for 2 h. The solid was collected by centrifugation, rinsing the filter cake with water
(60.1 kg) twice. The solid was recharged to a reactor with MTBE (133.6 kg) and stirred at 20°C
for 1 h. The solid was collected by centrifugation and washed with MTBE (44.4 kg), then dried at
40°C to give 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-pyrrolo[2,3-b]pyridine 25 (47.3
kg, 63.7% yield) as a white solid; HPLC, 99.2%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of imino-methyl-[1-[6-[(3R)-3-methylmorpholin-4-yl]-2-(1H-pyrrolo[2,3-b]pyridin-
4-yl)pyrimidin-4-yl]cyclopropyl]-oxo-λ6-sulfane 1; To a solution of [1-[2-chloro-6-[(3R)-3-
methylmorpholin-4-yl]pyrimidin-4-yl]cyclopropyl]-imino-methyl-oxo-λ6-sulfane
10
hydrochloride salt (61.6 kg, 1.0 equiv) in ethanol (806.1 kg) was added potassium carbonate
(73.6 kg, 3.3 equiv) and water (291.6 kg). The reaction mixture was degassed with nitrogen for 2
h, then 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-pyrrolo[2,3-b]pyridine 25 (53.5 kg,
1.3 equiv) and [1,1’-bis(diphenylphosphino)ferrocene]dichloropalladium(II) (2.6 kg, 0.022 equiv),
were added followed by a further 2 h of degassing. The reaction mixture was heated to ~80°C and
held for 3.5 h. The mixture was cooled to ~20°C and stirred for 5 h. The mixture was filtered to
remove the solids, the filter washed with ethanol (41.8 kg) and the filtrate concentrated under
vacuum down to 4 relative volumes, keeping the temperature below 50°C. The mixture was
extracted with ethyl acetate (367.6 kg) twice, and the combined organic layers washed with water
(69.1 kg). Silica thiol (55.3 kg) was added to the solution and stirred at 25°C for 8 h. The solid was
removed by filtration, and washed with ethylacetate (54.2 kg). The filtrate was concentrated under
vacuum down to 2.6 relative volumes, keeping the temperature below 50°C. 1-Butanol (80.7 kg)
was added and the mixture concentrated at ~70°C until 3.5 relative volumes were left. 1-Butanol
ACS Paragon Plus Environment
28

Page 29 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
(84.6 kg) was added and the volume again reduced by distillation to 2.5 relative volumes. Seed
crystals of imino-methyl-[1-[6-[(3R)-3-methylmorpholin-4-yl]-2-(1H-pyrrolo[2,3-b]pyridin-4-
yl)pyrimidin-4-yl]cyclopropyl]-oxo-λ6-sulfane 1 (37.0 g), were added at 75°C and the reaction
mixture was held for 2 h. The mixture was cooled to 20°C and n-heptane (170.0 kg) was added at
the rate of 40 kg/h. The solid was collected by filtration, washed with a mixture of 1-butanol and
n-heptane (10.6 kg and 26.7 kg respectively), and then a further wash with n-heptane (36.2 kg).
The solid was dried at 40°C to give imino-methyl-[1-[6-[(3R)-3-methylmorpholin-4-yl]-2-(1H-
pyrrolo[2,3-b]pyridin-4-yl)pyrimidin-4-yl]cyclopropyl]-oxo-λ6-sulfane 1 as a white solid (45.0
kg, 65.0% yield). Assay (HPLC) 99.9 %; Assay (NMR) 99 %wt/wt.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Optional rework of imino-methyl-[1-[6-[(3R)-3-methylmorpholin-4-yl]-2-(1H-pyrrolo[2,3-
b]pyridin-4-yl)pyrimidin-4-yl]cyclopropyl]-oxo-λ6-sulfane 1 to reduce metal levels if required;
imino-methyl-[1-[6-[(3R)-3-methylmorpholin-4-yl]-2-(1H-pyrrolo[2,3-b]pyridin-4-yl)pyrimidin-
4-yl]cyclopropyl]-oxo-λ6-sulfane 1 (45.0 kg) was dissolved in 1-butanol (184.0 kg), and heated to
45°C. Water (46.4 kg) was added and the mixture was cooled to 25°C. Silica thiol (5.0 kg) was
added and the reaction mixture was stirred for 8 h. The mixture was filtered to remove the solid
and the filter cake was washed with 1-butanol (46.1 kg). The filtrate was concentrated under
vacuum down to 2.6 relative volumes, keeping the temperature below 50°C. 1-butanol (73.7 kg)
was added and the mixture concentrated at ~70°C until 2.5 relative volumes were left. Seed crystals
of
imino-methyl-[1-[6-[(3R)-3-methylmorpholin-4-yl]-2-(1H-pyrrolo[2,3-b]pyridin-4-
yl)pyrimidin-4-yl]cyclopropyl]-oxo-λ6-sulfane 1 (30.0 g), were added at 75°C and the reaction
mixture was held for 2 h. The mixture was cooled to 20°C and n-heptane (141.9 kg) was added at
the rate of 40 kg/h. The solid was collected by filtration, washed with a mixture of 1-butanol and
n-heptane (9.3 kg and 22.4 kg respectively), and then a further wash with n-heptane (32.2 kg). The
ACS Paragon Plus Environment
29