Synthetic Route Design of AZD4635, an A2AR Antagonist

Article abstract of DOI:10.1021/acs.oprd.9b00171

The AstraZeneca approach to synthetic Route Design is exemplified through our AZD4635 chemical development program. The identification of possible new route concepts is presented, as well as their subsequent prioritization for practical exploration based on project objectives. Selected ideas were tested to demonstrate proof of concept for the bond formation strategy and, where successful, were fed into a decision tool based on key SELECTion principles.

Full text of DOI:10.1021/acs.oprd.9b00171

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC
Full Paper
Synthetic Route Design of AZD4635, an A2AR Antagonist
Mairi McCormick Littleson, Andrew D. Campbell, Adam Clarke, Mark Dow, Gareth J. Ensor, Matthew C.
Evans, Adam Herring, Bethany Jackson, Lucinda V. Jackson, Staffan Karlsson, David J. Klauber, Danny
Legg, Kevin Leslie, Stefan Moravcik, Chris D. Parsons, Thomas Oliver Ronson, and Rebecca E. Meadows
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00171 • Publication Date (Web): 25 Jun 2019
Downloaded from http://pubs.acs.org on June 26, 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 23
Organic Process Research & Development
1
2
3
4
5
6
7
Synthetic Route Design of AZD4635, an A2AR
Antagonist
8
9
Mairi M. Littleson,1* Andrew D. Campbell,1 Adam Clarke,1 Mark Dow,1 Gareth Ensor,1
Matthew C. Evans,1 Adam Herring,1 Bethany A. Jackson,1 Lucinda V. Jackson,1 Staffan
Karlsson,2 David J. Klauber,1 Danny H. Legg,3 Kevin W. Leslie,1 Štefan Moravčík,2 Chris D.
Parsons,3 Thomas O. Ronson1 and Rebecca E. Meadows.1
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
1Chemical Development, Pharmaceutical Technology and Development, AstraZeneca,
Macclesfield Campus, SK10 2NA
2Early Chemical Development, Pharmaceutical Sciences, IMED Biotech Unit, AstraZeneca,
Gothenburg, SE-431 83 Mölndal, Sweden
3Early Chemical Development, Pharmaceutical Sciences, IMED Biotech Unit, AstraZeneca,
Macclesfield Campus, SK10 2NA
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 23
1
2
3
4
For Table of Contents only
5
6
7
8
Route Design of AZD4635
Disconnection 1
7
6
5
4
3
2
Me
Disconnection 2
9
N
N
N
Route Design
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
1
Cl
SELECT criteria
N
NH2
F
Disconnection 3
Route Decision
ACS Paragon Plus Environment

Page 3 of 23
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
ABSTRACT: The AstraZeneca approach to synthetic Route Design is exemplified through
our AZD4635 chemical development programme. The identification of possible new route
concepts is presented, as well as their subsequent prioritization for practical exploration based
on project objectives. Selected ideas were tested to demonstrate proof of concept for the bond
formation strategy, and where successful, were fed into a decision tool based on key
SELECTion principles.
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: Route Design, Process Design, SELECT criteria, Route Decision
INTRODUCTION
Within AstraZeneca, effective, ambitious Route Design is viewed as integral to the
development process, enabling early identification of the best synthetic route for commercial
supply. We define Route Design as the identification of our bond formation strategy. Further
development of the overall delivery strategy, such as the selection of reagents and conditions
(Process Design and Optimization), are a consideration but not a priority of the Route Design
exercise. The culture within AstraZeneca promoting Route Design as the most value adding
activity in the development process has led to tangible benefits, including savings in supply
chain costs, waste reduction, focused application of resource, and reduced regulatory
challenge.1
AZD4635 (1) is an A2AR antagonist currently in clinical trials for the treatment of solid
tumours.2 To date, the original discovery route has been used to meet clinical demands for
AZD4635 following development to enable appropriate scale-up. This route is concise and
operationally robust, affording 1 in ca. 30% yield over five steps, and has been used to produce
6.5 kg of the active pharmaceutical ingredient (API) (Scheme 1).3,4,5
H2N
N
NH
Br
N
HO
OH
O
N
H2CO3
N
3
H2N
NH
AcOH
1) Br2, pyridine, MeOH
N
NH2
N
NH2
2) Na2S2O3, K2CO3
65-67% yield
56-63% yield
F
4
5
F
F
2
Me
Me
[(1,5-COD)(OMe)Ir]2
DTBBP, B2Pin2
N
N
Me
Me
Me
O
n-heptane
76% yield
Cl
Cl
B
N
1) Pd(OAc)2, Amphos
K2CO3, THF/H2O
Recrystallization
6
DMSO/i-PrOH
O
N
N
AZD4635
API
Cl
N
83% yield
2) EtOH crystallization
Me
Me
 Short, convergent synthesis
 Produced 6.5 kg of the API
 28% overall yield
84-85% yield
NH2
7
F
1
Scheme 1. Developed Route to AZD4635
The key feature of this route is the regioselective preparation of the 3-amino-1,2,4-triazine6
core 4 through condensation of 2 and amino guanidine 3. Subsequent bromination gave access
to 5 with Suzuki-Miyaura coupling of BPin 7, obtained through iridium catalyzed borylation
of pyridine 6, delivering AZD4635. The bromination step (4 to 5) afforded a moderate isolated
of 5 (56-63%).5 Furthermore, the bromination conditions were identified as being incompatible
with tantalum reactor components,5 and a previous manufacture of the API had to be halted
due to reactor corrosion. Although manufacture could be carried out using vessels without
metal components, to reduce risk, allow greater flexibility for manufacture, and potentially
allow for improvements in overall yield, step count, etc., a second-generation synthesis was
desirable for future API deliveries. This prompted a Route Design exercise assessing
alternative strategies towards AZD4635, culminating in a Route Decision for future supply of
API.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 23
1
2
3
4
5
RESULTS AND DISCUSSION
6
7
8
9
Our Route Design approach begins with idea generation supported by literature searching and
the use of retrosynthesis software.7 We sought to be ambitious in our proposed disconnections
and identified several synthetic strategies towards the API (Scheme 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
Me
 Established, robust synthesis of
bi-aryl triazene fragment
 Alternative means of forging
C6-C7 desirable
N
M
Cl
Disconnection 1
Me
X
N
N
N
1
Me
Disconnection 2
9
NH2
8
7
10N
11
N
N 1
F
6
5
Cl
N2
3
2
X
X
H2N
H2N
12
18
Cl
N
17
16
4 N
NH2
13
14
NH2
F
 Highly convergent strategy
 Sequential or chemoselective
bond formation
3
15
X
N
F
N
Disconnection 3
 Final step triazine formation
 Step economic
 Potential regiochemical issues
X
N
NH2
Me
M
N
M
Cl
F
Scheme 2. Key disconnections identified
The synthetic strategies proposed were grouped by key bond disconnections: Disconnections
1 and 3 around the triazine ring, and disconnection 2 of the ring itself. These strategies were
then prioritized taking into consideration which approaches were viewed as having a higher
likelihood of success based on literature precedent and which were seen as the most value
adding with respect to cost, manufacturability, and control (risk/reward). Step count was used
as a surrogate for cost at the proof of concept stage, and strategies with a perceived high level
of risk due to limited or the absence of literature precedent were prioritized where success
would be highly beneficial to the project (high risk, high reward). Disconnections dissecting
the amino guanidine unit were avoided as they were not perceived to offer improvement of the
original route.8
Proof of Concept Work. This strategy could then allow the selection of the most promising
routes for further investigation, with specific focus on establishing a proof of concept for key
bond formations. In parallel to this, further Process Design activities were also commenced on
the original route, with the aim of identifying milder bromination conditions for this stage.
Disconnection 1. The C6-C7 disconnection, which was used in the original synthesis,5 merited
further investigation. The robust, efficient synthesis of amino triazine 4 made this
disconnection an attractive prospect and therefore alternative means of activating C6 were
investigated. (Scheme 3).
ACS Paragon Plus Environment

Page 5 of 23
Organic Process Research & Development
1
2
3
4
Disconnection 1
a)
C-H activation approach
Me
Me
Me
5
6
7
8
N
N
N
N
N
N
N
N
N
7
N
6
or
Cl
F
N
NH2
SMe
Cl
Cl
Coupling partners
Cl
BPin
7
9
N
NH2
F
F
8
4
1
9
No reactivity observed
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
b)
Phosphonium salt coupling
Li
N
N
Me
N
Me
N
NR2
11
N
Tf2O, PPh3
NEt3
F
OTf
API
Cl
Cl
PPh3
6
10
Scheme 3. C6-C7 disconnection proof of concept work
To avoid pre-functionalization of C6, a C-H activation strategy could be a powerful and atom
economic means of forging the C6-C7 bond, removing a step from the original route (Scheme
3a).9 Although this was viewed as a higher risk route strategy, due to a lack of literature
precedent for the transformation on these substrates, the opportunity to exploit high-throughput
experimentation (HTE) techniques to rapidly assess this option made its investigation viable.
Thiomethyl analogue 8 was prepared to alleviate concerns over the free NH2 motif in the C-H
activation. A range of conditions were assessed on both 4 and 8, using aryl BPin 7 and aryl
chloride 9 as coupling partners.10 Disappointingly, no reactivity was observed in this screen,
and as such this tactic of C6-C7 bond formation could be eliminated from further investigations.
Reports detailing the use of phosphonium salts as pseudo halogens for reaction with
nucleophiles inspired us to target phosphonium salt 10 for subsequent reaction with lithiated
species 11 (Scheme 3b).11 This strategy also gave a reduced step count from the original route,
and was hoped to be a successful, high yielding means of forging the C6-C7 bond. 2-Halo
pyridines are known to be competent substrates for phosphonium salt formation,11 however
attempts to synthesise 10 from pyridine 6 did not afford the desired product, rather a mixture
of unidentifiable by-products, and this approach was not taken further.12
Disconnection 2. The N1-C6, N4-C5 disconnection was proposed as an efficient means of
assembling the fully substituted amino triazine core in a single step (Scheme 4).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 23
1
2
3
4
Disconnection 2
a) Flow approach
Me
Me
Me
O
O
MeOH
5
6
7
8
N
Me
Cl
N
N 1
N
20 °C, 4 mins
6
5
O
Cl
N
O
N
N
Cl
Cl
4 N
NH2
ZnCl
R
)
Cl
Me
16
14
R = Im (
15
or OMe ( )
0.5 M in THF
1
13
0.5 M in THF
F
9
12
Low conversion
F
F
F
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
By-product formed in batch
b) Lithiation approach to 19
c) Chlorination and cyclization
OEt
Me
Me
Me
O
H2N
H2N
Me
NH
N
H2CO3
N
18
LiHMDS
N
F
N
3
Cl
O
NH
NCS
Cl
Cl
API
6
Cl
NaI, EtOH, air
10-30% yield
THF
PTSA
Me
Cl
78% yield
CH2Cl2
96% yield
O
O
rr 9:1 (1:23)
F
F
F
Low yielding cyclization
20
17
19
19
d) Cu oxidation and condensation
Me
N
Me
Me
H2N
NH
HCl
N
N
22
Cu(OAc)2·H2O
H2N
NH
N
N
NH2
O
O
API
Cl
Cl
Cl
K2CO3
i-PrOH, TFA, 70 °C
32% yield
rr 4:1 (1:23)
N
MeCN, air
77% yield
O
23
F
F
F
19
21
API regioisomer
e) Alkyne addition approach
Cl
N
Me
H2N
25
NH
H2CO3
Me
Me
N
F
3
H2N
NH
Pd(PPh3)2Cl2, CuI
NaI, Chloramine T
API
N
NEt3, 1,4-dioxane
80 °C
90% yield
THF/H2O, 80 °C
68% yield
I2, DMSO, 80 °C
50% yield
rr 3:1 (1:23)
Cl
Cl
I
BPin
 Condensation via oxidation
 Unoptimized conditions
7
24
26
F
Scheme 4. Disconnection 2 proof of concept work. rr = regioisomeric ratio (1:23)
We investigated the synthesis of a biaryl fragment comprising of the pyridyl and p-fluoro
benzene rings, which could then be reacted with the amino guanidine 3 to deliver the API.
We targeted mono-ketone 12, envisaging a step-wise condensation with amino guanidine 3,
activation of the methylene, and final cyclization to furnish the API. Although linear in nature,
this approach would allow access to the API in one step less than the original route and was
perceived to have a high likelihood of success. Attempts to synthesize mono-ketone 12 from
reaction of acyl imidazole 14 and zinc species 13 had resulted in over reaction due to the
basicity of the zinc species, forming an impurity consistent with di-ketone 16. We saw an
opportunity to utilize flow chemistry to circumvent this issue by avoiding high concentrations
of the reactive species in the presence of the desired product (Scheme 4a). Disappointingly,
under these conditions very low conversion to the desired product was observed, and the
comparative reaction using methyl ester 15 also gave no conversion to 12. Alternatively, mono-
ketone 19 was targeted. Synthesis of 19 proved facile; ester 18 was reacted with the lithium
anion of pyridine 17, affording mono-ketone 19 in 78% yield (Scheme 4b).13
Initial attempts to access a cyclization precursor from 19 focused around the synthesis of α-
chloro ketone 20 (Scheme 4c). Treatment of 19 with N-chlorosuccinimide (NCS) gave
chlorinated compound 20 cleanly and selectively. Cyclization using catalytic NaI and air as
oxidant typically afforded AZD4635 in yields of 10–30% and in 9:1 selectivity for the desired
ACS Paragon Plus Environment

Page 7 of 23
Organic Process Research & Development
1
2
3
4
5
6
regioisomer. The low yield was due to the formation of several by-products resulting from non-
ring closed intermediates and competing dimerization. Attempts to improve the selectivity of
this reaction were unsuccessful, and therefore alternative means of activating C6 were
investigated.
7
8
9
Di-ketone 21 was then targeted as a means of bringing C5 and C6 in at the correct oxidation
state, despite initial concerns regarding the regioselectivity of the proposed condensation
reaction (Scheme 4d).8 Initial screening for the oxidation of mono-ketone 19 identified a hit
using Cu(OAc)2·H2O under air atmosphere. The di-ketone could then be condensed with the
amino guanidine 22 to give the API in promising yield of 32% with 4:1 regioselectivity
meriting further development vide infra and providing proof of concept for this route.
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
Alternatively, alkyne 26 was viewed as a potential API precursor, in a relatively high risk, but
potentially extremely atom efficient strategy (Scheme 4e). Literature precedent for in situ
oxidation of alkynes and subsequent condensation with diamine units,14 as well as the
promising results from the di-ketone route (Scheme 4d), gave us optimism for this approach.
Bi-aryl alkyne 26 proved a challenging intermediate to access efficiently. In principle, 26 could
be formed through a Sonogashira reaction, disconnecting either the C6-C7 bond or C5-C13 bond
depending on the coupling partners chosen. As BPin 7 was available in-house, deborylative
iodination using NaI and Chloramine T was carried out successfully,15 affording iodo-pyridyl
species 24 in 68% yield. Sonogashira coupling of 24 with aryl alkyne 25 proceeded in 90%
yield, giving access to our desired alkyne 26. Under non-optimized conditions, oxidation of the
alkyne and subsequent condensation with amino guanidine 3 afforded the API in 50% yield
and 3:1 regioisomeric ratio, in favour of the desired regioisomer 1. Although this route
delivered the API successfully, it was not viewed as favourably as the route from mono-ketone
19 due to the inefficient access to alkyne 26, and the lack of any obvious way to improve its
synthesis.
The most promising route strategy arising from this disconnection was therefore considered to
be the formation of mono-ketone 19, followed by oxidation to di-ketone 21, and final
condensation with amino guanidine 22 to form the API.
Disconnection 3. Bringing in the amino triazine unit intact was an attractive prospect regarding
the convergency and efficiency of the synthesis. This strategy would rely on chemoselective
bond formation between C6 and C7, and C5 and C13 (Scheme 5).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 23
1
2
3
Disconnection 3
a) Catellani approach
4
5
6
7
8
9
Me
N
N
I
N
Br
N
N
N
N
N
NBS
N
TBAC, K2CO3
MeCN
80% yield
7
N
NH2
N
NH2
N
6
N
NH2
N
NH2
Cl
N
27
28
27
29
5
N
NH2
13
 No iodination observed
 Several conditions tried
1
F
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
Competition Experiment
Br
N
N
Me
Me
Me
N
N
N
NH2
 Selectivity shown for triazine
coupling
 Suggests desired selectvity for
Catellani is possible
Pd(OAc)2
29
N
N
N
+
Cl
+
Cl
Br
APhos, K2CO3
THF/H2O, 80 °C
Cl
BPin
N
NH2
F
7
F
87:13
30
31
32
Me
Br
N
N
Br
N
N
 Significant Suzuki coupling observed
 No product formation
+
+
API
Pd, ligand
base, solvent, 
NH2
F
Cl
BPin
29
30
7
b) Sequential cross coupling approach
Me
Me
N
7
N
Br
N
Br
N
H2O2
Cl
BPin
N
N
N
Cl
N
Pd(OAc)2
APhos, K2CO3
THF/H2O, 80 °C
AcOH
63% yield
N
NH2
HO
N
NH2
HO
34
N
NH2
33
29
Low conversion to 34
Scheme 5. Disconnection 3 proof of concept work
The Catellani reaction was viewed as a challenging, yet potentially rewarding means of
developing a highly convergent synthesis of the API (Scheme 5a).16 Disconnection of the
triaryl motif afforded three proposed fragments; iodo 28, known BPin 7, and aryl bromide 30.
Concerns were raised surrounding the use of the amino triazine core which cast doubt on the
viability of this chemistry with respect to our API. It was unclear how the presence of the ortho-
nitrogen of the desired iodo-triazine would affect the reaction, as the Catellani reaction
typically requires a substituent ortho to the iodide, known as the ortho-effect.17
Disappointingly, we were unable to access the desired iodo 28 from iodination of unsubstituted
amino triazine 27. Catellani type reactions have largely been limited to aryl iodides, the “aryl
iodide constraint”,18 however there are literature examples of aryl triflates19 and bromides.18
Bromination of 27 proceeded smoothly, and therefore we considered the use of bromo 29 in
the Catellani reaction. To cleanly access our API, we required chemoselective oxidative
addition to bromo 29 in the presence of 30.18 To give an indication of whether this approach
would be viable, we carried out a competitive Suzuki-Miyaura reaction using 29 and 30. Under
the conditions utilized for the Suzuki-Miyaura coupling to prepare AZD4635 in the original
route, the coupling was largely selective for the triazine bromide 29, affording 31 and 32 in
60% and 9% conversion respectively (ratio ca. 6.5:1). Encouraged by this result, we
investigated the desired Catellani reaction with bromo 29, bromo 30, and BPin 7.
Disappointingly, from a screen of typical Catellani conditions,18 we observed no formation of
the desired API.10 At this juncture we considered the significant amount of screening likely to
be required to develop conditions making the Catellani a competitive route with regard to the
alternative strategies being developed, and therefore we chose not to progress with the Catellani
based approach.
ACS Paragon Plus Environment

Page 9 of 23
Organic Process Research & Development
1
2
3
4
5
6
An analogous method was also envisaged where C6-C7 and C5-C13 could be formed through
sequential cross couplings (Scheme 5b). We proposed that access to bromo/hydroxy compound
33 could enable coupling with the required pyridyl nucleophile and, following conversion of
the hydroxy moiety to a triflate, enable coupling with a p-fluoro aryl nucleophile.
7
8
9
Key precursor 33 was prepared by oxidation of bromo 29 with H2O2.20 We then used the
Suzuki-Miyaura conditions developed for the preparation of AZD4635 to carry out the
coupling. Disappointingly, moderate conversion to product 34 was observed in this reaction,
ca. 40%, with the formation of several unidentified by-products. Clean isolation of the Suzuki-
Miyaura product was challenging and, due to time constraints, this approach was not
investigated further.
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
Process Design and Optimization of Selected Route Strategies. Following these proof of
concept studies, the most promising new strategy was identified as the di-ketone condensation
route (Scheme 4d) and therefore Process Design activities were initiated to develop this route
option.
The formation of mono-ketone 19 was further investigated. It was found that the
regioselectivity of deprotonation was both base and solvent dependent (Table 1).
Table 1. Base and solvent dependence of mono-ketone 19 formationa
O
Me
O
Me
O
N
F
N
Base
N
+
OEt
Cl
Solvent
F
N
Cl
Cl
Me
F
O
Cl
Me
O
F
F
17
18
19
35
36
Entry
Base
Solvent
Time (h)
19c
89
33
0
35c
1
36c
5
1
2
3
4
5
6
7
8
9
LiHMDS
LiHMDS
LiHMDS
LiHMDS
LiHMDS
NaHMDS
KHMDS
KHMDS
KHMDS
THF
4
4
4
4
4
4
4
4
4
2-MeTHFb
TBME
34
65
41
43
58
33
41
76
25
0
Toluened
Hexane
THF
0
0
0
0
31
29
18
7
3
THF
1
2-MeTHF
TBME
1
7
aReactions were carried out using 17 (7.06 mmol, 1.00 equiv), 18 (7.08 mmol, 1.00 equiv), base (1.00 M, 16.0 mmol, 2.30
equiv) at 20 °C. bLiHMDS 0.50 M in 2-MeTHF. cArea% calculated by HPLC at 220 nm. dArea% of toluene subtracted from
chromatograph.
The use of less polar solvents than THF led to a reversal of selectivity, affording regioisomer
35 rather than the desired product 19 (entries 3–5). 2-MeTHF gave equal amounts of 19 and
35, and an increase in double addition product 36 (entry 2). Change of the HMDS counterion
also gave a decrease in selectivity, with low conversion to the desired product (entries 6–9).
These effects are consistent with an interaction between the pyridyl nitrogen and the
counterion, facilitating the reaction at the less acidic 2–methyl group.21 Therefore, we retained
the original conditions of LiHMDS in THF.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 23
1
2
3
4
5
6
Diketone 21 was originally accessed through Cu-mediated oxidation of 19 under air (Scheme
4d). We sought to improve reaction understanding and assess the overall sequence as a
potential route for API manufacture (Scheme 6).
7
8
9
a) Optimized Cu-oxidation
Diketone development
Me
Me
Me
Me
N
N
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
Cu(OAc)2
O
O
H2N
H2N
O
O
Cl
Cl
NH
Cl
Cl
TMEDA, K2CO3
MeCN, air
80% yield
+
API
O
NH
O
F
F
F
F
21
19
19
21
b) Diazene intermediate and oxidation
Me
Me
Me
Me
N
NHBoc
NBoc
N
N
NH2
NH
TFA
CH2Cl2
LiHMDS
then
N
18
O
O
Cl
Cl
Cl
then
84% yield
Boc
Me
Cl
O
O
N
N
39
17
38
21
37
Boc
F
F
F
53% yield
Not observed
 Unexpected oxidation
 Facile formation of di-ketone
c) Flow synthesis of 38
d) Condensation with amino guanidine
Me
Me
17
and
18
Reactor 1
40 °C, 10 mins
H2N
NH
N
NHBoc
NBoc
Reactor 2
- 40 °C, 20 sec
N
HCl
22
1 M in THF
O
H2N
NH
Cl
Cl
API
19
Zn(OTf)2
rr 94:6 (1:23)
LiHMDS
1 M in THF
O
O
F
21
F
38
63% yield
37
0.5 M in THF
Scheme 6. Development of the di-ketone condensation route. rr = regisomeric ratio
We observed that the outcome of the oxidation was highly dependent on the agitation of the
reaction mixture and concluded that this was a result of variable oxygen mass transfer.
Optimized conditions of 5 mol% Cu(OAc)2 in MeCN, with tetramethylethylenediamine
(TMEDA), K2CO3 and bubbling air through the reaction gave full conversion of the starting
material in 16 h, and 21 in 80% isolated yield (Scheme 6a).
Due to potential safety concerns regarding the atmospheric oxidation,22 we assessed whether
the reaction could run efficiently under an atmosphere of 5% oxygen. Disappointingly, it was
found that the oxidation only progressed to ca. 20% conversion after 20 h under these
conditions. Running the reaction under air in a water/MeCN (9:1 v/v) solvent mixture to reduce
flammability did not give any conversion to the desired product. There are literature reports of
using (NMP), a high flash point solvent (90 °C), to allow the safe use of air as an oxidant,
however in this solvent conversion to the desired product was very low ca. 5%.23
Given these safety concerns, alternative routes from 19 were investigated. Quenching of the
lithium enolate of 19 with diazene 37 afforded intermediate 38 (Scheme 6b). When
deprotection of 38 was attempted under acidic conditions, with the hope of reacting the
deprotected product 39 with urea equivalents such as cyanamide, the unexpected formation of
di-ketone 21 was observed. This was not dependant on atmospheric oxygen as equivalent
results were obtained when the reaction was carried out under inert conditions. Although the
ACS Paragon Plus Environment

Page 11 of 23
Organic Process Research & Development
1
2
3
4
5
specific mechanistic details of this transformation remain unclear, we assessed this route in
flow as an alternative means of safely accessing di-ketone 21 on scale (Scheme 6c).
6
7
8
9
Assessment of flow conditions towards the mono-ketone (19) showed that although higher
temperatures enabled shorter residence times, they also increased formation of a by-product
resulting from dimerization. Therefore, a residence time of 10 minutes at 40 °C was selected.
The quench with diazene 37 had been carried out successfully at –70 °C in batch, however this
was not compatible with the flow process as the 0.5 M solution of 37 began to crystallize at
temperatures lower than –50 °C. A residence time of 20 seconds at –40 °C was found to be
optimal, and this process gave 38 in 63% yield after work-up and crystallization. Deprotection
to di-ketone 21 could then be carried out in batch. With a means of accessing di-ketone 21 in
hand, we focused on the key condensation with amino guanidine 22 (Scheme 6d).
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
In initial experiments, we screened a range of conditions to assess di-ketone conversion, by-
product formation, and regioselectivity.10 The best conditions identified at this point were 1.1
equiv 22 in H2O/i-PrOH at reflux, affording AZD4635 in 50% isolated yield and 87:13
regioselectivity. Several by-products resulting from condensation of the di-ketone with two
equivalents of amino guanidine, or incomplete condensation were identified, and further
optimization sought to reduce these. We found that guanidine salts of stronger acids i.e. HCl
gave improved regioisomeric ratios as did reduced water levels. Where acid additives were
added (MeSO3H or H3PO4), higher charges of the additive also improved regioselectivity,
however the higher acid charge resulted in reduced yields arising from degradation.10
This prompted us to assess Lewis acids as an alternative means of activating the reaction. A
screen of Lewis acid mediators was carried out,10 and the reactions showing promising (> 9:1)
regioselectivity were considered for further development. Zn(OTf)2 gave a regioisomeric ratio
of 94:6 in favour of the desired regioisomer, however conversion to the API was incomplete.
Due to time constraints, when the Route Decision was made, these conditions had not been
progressed further, however we had obtained proof of concept and were confident the reaction
could be optimized. Further studies showed that the unwanted regioisomer 23 could be purged
to 0.04% by slurrying in hot toluene, which gave us further confidence in the potential of this
route.
Overall, this three-step route (mono-ketone 19 formation, oxidation to 21, and condensation
with amino guanidine) delivered the API in 30% yield and high purity, and was therefore
competitive with the current route, despite the limited optimization carried out. Although the
oxidation of 19 to 21 gave concerns with regard to process safety on scale up, we were
confident that safe, efficient conditions could be established. We considered running the
oxidation in flow or the use of a chemical oxidant as safer alternatives to be investigated should
the route be progressed, and the potential for these improvements was taken into account in the
Route Decision.
During the Route Design exercise, Process Design on the original bromination was carried out
with the aim of identifying milder conditions for this transformation. Both bromination and
iodination conditions for 4 were assessed via a HTE format.10 A hit was identified from this
screen and, following further optimization, new conditions were identified. N-
bromosuccinimide (NBS) and K2CO3 in MeCN displayed 98% conversion of the starting
material (Scheme 7).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 23
1
2
3
New bromination conditions
4
5
6
7
8
9
N
N
Br
N
N
NBS
K2CO3
MeCN
98% conversion
N
NH2
N
NH2
4
5
F
F
 HTS screen
 Iodination/bromination conditions
 Hit identified
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 7. Bromination hit from HTE screening
As the initial bromination hit required stringent inertion and was capricious outside of a
glovebox, a second round of HTE was carried out to develop these conditions. The screen was
carried out in a 96-well plate format, looking at a wide range of bromination conditions (Table
2).
Table 2. Optimization of bromination hita
N
Br
N
N
N
Br source
base, additive
solvent
N
NH2
N
NH2
4
5
F
F
Br source
(equiv.)
Base
(equiv.)
Additive
(equiv.)
5 (%)b
Entry
Solvent
NBS
(1.2)
1
2
3
4
5
6
7
8
9
DBU (2)
-
MeCN
EtOAc
EtOAc
MeCN
MeCN
MeCN
PrCN
71
82
74
67
78
NBS
(1.2)
DBU (2)
DBU (2)
-
DBH
(0.6)
-
NBS
(1.2)
TBACl (2)
K2CO3 (2)
K2CO3 (2)
K2CO3 (2)
K2CO3 (2)
K2CO3 (2)
-
NBS
(1.2)
-
quant.
(64c)
NBS
(1.2)
TBACl (1)
TBACl (1)
TBACl (1)
TBACl (1)
NBS
(1.2)
99
96
86
DBH
(0.6)
MeCN
MeCN
Br2 (1.2)
aReactions were carried out using 0.1 mmol 4, at 25 °C in a glovebox. HPLC conversion at 20 h, m-terphenyl (0.04 M in
b
PhMe) used as internal standard. cReaction carried out on 1.05 mmol 4, isolated yield.
The addition of tetrabutylammonium chloride (TBACl) was found to be beneficial (presumably
generating BrCl as the active brominating agent in situ).24 NBS was typically the best
brominating agent, followed by 1,3-dibromo-5,5-dimethylhydantoin (DBH), with bromine the
least effective. NBS in combination with K2CO3 and TBACl gave full conversion in MeCN
and could reliably be performed outside a glovebox (Entry 6). At this point the work-up and
isolation of 5 was unoptimized, affording an isolated yield of 64%. The cause of the low
isolated yield was unclear however we were confident that with further development this could
be improved. These milder bromination conditions did not display the corrosive properties of
the previous reaction (Scheme 1),5 and therefore rendered the original route significantly more
attractive with respect to further campaigns.
Route Decision. At the time of the scheduled route decision, the new route work had not
undergone significant development, however we undertook the decision process considering
the potential for the improvement of the new routes through further process development. We
defined our parameters for assessment using SELECT criteria; the Safety of the route, the
ACS Paragon Plus Environment

Page 13 of 23
Organic Process Research & Development
1
2
3
4
5
6
Environmental impact of the route, Legal considerations, Economic factors, the Control of the
route, and the potential Throughput of the route.25 These parameters were considered in turn
for the routes assessed, and a benchmark set i.e. all parameters in the current route were scored
5.
7
8
9
When all potential new routes were assessed against each other, the di-ketone condensation
route (Scheme 6), furthermore known as New Route One (NR1), was considered the most
promising and lowest risk of the new approaches. Although the new bromination conditions
had not undergone significant development, the bromination step was embedded in an
otherwise reliable and well-developed route.5 Therefore, a Pugh matrix26 was used to compare
NR1 and the original route, incorporating the newly developed bromination conditions, with
the original route set as the benchmark (Figure 1).
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
Safety
7
Original route
6
NR 1
5
Throughput
Environment
4
3
2
1
Legal
Control
Economics
Figure 1. Pugh matrix scoring the original route (new bromination conditions) and NR1
Consideration of the SELECT criteria surrounding the original route identified risk regarding
the inclusion of the new bromination process and therefore potentially new impurities, and
possible problematic off-gassing from the borylation stage (H2). The environmental impact of
the Ir catalyst and the relatively low atom economy (halogenation/borylation stages) were
considered. At the time of the route decision, we had limited understanding of the new
bromination reaction, including its mechanism, inertion requirements, and concerns
surrounding product isolation. However, the route historically delivered high quality material,
in good overall yield (35% excluding final recrystallization), and therefore was seen as a robust
means of API delivery.
As previously mentioned, NR1 had safety concerns surrounding the use of atmospheric oxygen
and with respect to the purging of impurities which had undergone little development. Each
stage required high dilution (≥ 30 L/kg limiting reactant), however NR1 had excellent atom
economy, and was a step shorter than the original route. However, the formation of the API
regioisomer in the final bond-forming step of the process, and limited purge data available
represented a significant risk to the control strategy for this route going forward. The route
supplied high quality material in good overall yield (30% excluding final recrystallization),
however had not yet been assessed beyond gram scale.
A direct comparison of the key parameters is shown in Table 3.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 23
1
2
3
4
5
6
7
Table 3. Direct Comparison of Key Parameters
8
9
Original Route (New Bromination)
Parameter
NR1
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
Step Count
Yield
3 linear
Predicted 30%
High
4 (3 linear)
Predicted 35%
Atom Economy
Bromination and Borylation
CONCLUSION
When all factors were considered, the decision was taken to progress the original route. Despite
the fact that the new routes were not progressed, the Route Design exercise could still be
considered a success. The strategic framework of an AZ Route Design programme was used to
ensure the route decision taken was as informed as possible. The new routes investigated were
identified and the prioritization of these routes was based on our ideal of ambitious Route
Design, including strategies which were viewed as “high risk, high reward” such as the
Catellani approach. Despite the short time frame available, the new route work was carried out
to achieve proof of concept of novel bond disconnections and was successful in delivering new
routes to contend with the original process. The information generated enabled structured
decision making to confidently make the best informed Route Decision.
EXPERIMENTAL METHODS
General. All reactions were carried out in dry vessels under an atmosphere of dry nitrogen
unless otherwise specified. All reagents and solvents were used as received without further
purification unless specified. NMR spectra were recorded on a Bruker 500 MHz or 400 MHz
spectrometer. Accurate mass spectra analyses were recorded on a Waters LC-MS ZQ 2000
system in positive electrospray ionization mode (ESI). Reactions carried out in microwave
tubes were brought to the desired reaction temperature in a heating block.
HPLC Methods. Flow synthesis of 12. Samples for HPLC analysis (conversion to product)
were prepared through the addition of 1:1 (v/v) MeCN:H2O (1 mL) to a sample of the reaction
following quench with MeOH. HPLC was obtained on an Agilent 1100, using a Phenomenex
Kinetex C18 column. Analysis was performed using a gradient method, eluting with 5% of a
1% (v/v) aqueous TFA solution, 5-90% MeCN in H2O over 4.8 minutes at a flow rate of 1.2
mL/min and a wavelength of 220 nm.
Mono-ketone formation screening (Table 1). Samples for HPLC analysis (20 µL aliquot of
crude reaction) were diluted into a 10% solution of citric acid in 1:1 (v/v) MeCN/H2O (1.5
mL), injection volume 3 µL. HPLC was obtained on an Agilent 1100, using a Acquity BEH
Phenyl column. Analysis was performed using a gradient method, eluting with 3% of a 1%
(v/v) aqueous TFA in 5-90% MeCN in H2O over 6.2 minutes at a flow rate of 2.0 mL/min and
a wavelength of 220 nm.
AZD4635 regioisomer determination. Samples for HPLC analysis (20 µL aliquot of crude
reaction) were diluted into 1:1 (v/v) MeCN/H2O (1.5 mL), injection volume 3 µL. HPLC was
obtained on an Agilent 1100, using a Acquity BEH C18 column (40°C). Analysis was
performed using a gradient method, eluting with 0.003% (v/v) TFA in MeCN and 0.003% (v/v)
ACS Paragon Plus Environment

Page 15 of 23
Organic Process Research & Development
1
2
3
4
5
TFA in H2O over 17.3 minutes (5%-90%), then held to 23.5 minutes at a flow rate of 2.0
mL/min and a wavelength of 271 nm.
Halogenation screening (Scheme 7). Samples for HPLC analysis were prepared by taking a 10
µL aliquot of the reaction mixture and diluting into 4:1 (v/v) MeCN/H2O (1.5 mL), 1.0 µL
injection volume. HPLC was obtained on an Agilent 1100, using a Phenomenex Kinetex C18
column. Analysis was performed using a gradient method, eluting with 5-95% of a 0.03% (v/v)
TFA/MeCN solution in a 0.03% (v/v) aqueous TFA solution over 5 minutes at a flow rate of
1.2 mL/min and a wavelength of 220 nm.
Bromination screening (Table 2). Samples for HPLC analysis were prepared through the
addition of DMSO (200 µL) to the reaction and stirring for 10 minutes. A 10 µL aliquot was
then taken and diluted into 4:1 (v/v) MeCN/H2O (1.5 mL), 0.7 µL injection volume. HPLC
was obtained on an Agilent 1100, using a Phenomenex Kinetex C18 column. Analysis was
performed using a gradient method, eluting with 5-95% of a 0.03% (v/v) TFA/MeCN solution
in a 0.03% (v/v) aqueous TFA solution over 5 minutes at a flow rate of 1.2 mL/min and a
wavelength of 220 nm.
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
Characterization data of AZD4635 (1): 1H NMR (500 MHz, DMSO-d6): δ 7.72 (br s, 2H),
7.51-7.46 (m, 2H), 7.30-7.24 (m, 3H), 7.15 (s, 1H), 2.40 (s, 3H); 13C NMR (126 MHz, DMSO-
d6): δ 163.3 (d, 1JC-F = 248.5 Hz), 161.9, 159.1, 155.9, 149.1, 147.9, 145.4, 131.9 (d, 3JC-F = 8.8
Hz), 131.9, 122.0, 120.6, 115.5 (d, 2JC-F = 21.9 Hz), 23.60; 19F NMR (470 MHz, DMSO-d6): δ
˗109.96; HRMS: exact mass calculated for [M+H]+ (C15H12N5ClF) requires m/z 316.0765,
found m/z 316.0766.
Synthesis of 8. Aminothiourea (5.00 g, 54.9 mmol, 1.00 equiv) and EtOH (55 mL) were charged
to a 100 mL flask and iodomethane (3.4 mL, 55 mmol, 1.00 equiv) added in one portion. The
reaction was brought to 75 °C for 5 h. The reaction was allowed to cool to room temperature
and the solvent removed in vacuo to afford a pale yellow solid. The material was washed with
MTBE (30 mL) and filtered under vacuum until dry to afford 3-amino-2-methyl-isothiourea
hydroiodide as an off-white solid (12.32 g, 52.86 mmol, 96%) which was used without further
purification. 1-(4-Fluorophenyl)-2,2-dihydroxy-ethanone (1.00 g, 5.88 mmol, 1.00 equiv), 3-
amino-2-methyl-isothiourea hydroiodide (1.51 g, 6.47 mmol, 1.10 equiv), and acetic acid (12
mL) were charged to a flask and the reaction brought to 75 °C for 3 h. The reaction was allowed
to cool to room temperature, diluted with water (200 mL) and extracted with EtOAc (4 x 100
mL). The organics were combined, dried over Na2SO4, filtered, and evaporated to afford a dark
brown residue. The material was purified by flash silica column chromatography, eluent 20%
EtOAc/iso-hexane, to afford a dark orange solid. The material was then dissolved in minimal
CH2Cl2, and iso-hexane added until a precipitate formed. The solid was filtered and dried in
1
vacuo to afford triazine 8 as a pale orange solid (708 mg, 3.20 mmol, 54%). H NMR (400
MHz, DMSO-d6): δ 9.82 (s, 1H), 8.45-8.36 (m, 2H), 7.50-7.43 (m, 2H), 2.68 (s, 3H); 13C NMR
(101 MHz, DMSO-d6): δ 172.4, 165.0 (d, 1JC-F = 251.8 Hz), 153.1, 142.6, 130.6 (d, 3JC-F = 9.3
Hz), 129.3 (d, 4JC-F = 3.0 Hz), 116.5 (d, 2JC-F = 22.0 Hz), 13.2; 19F NMR (470 MHz, DMSO-
d6): δ ˗106.56; HRMS: exact mass calculated for [M+H]+ (C10H9N3FS) requires m/z 222.0501,
found m/z 222.0504.
Flow synthesis of 12. 2-Chloro-6-methyl-pyridine-4-carboxylic acid (2.57 g, 15.0 mmol, 1.00
equiv), 1,1’-carbonyldimidazole (2.55 g, 15.4 mmol, 1.03 equiv), and THF (24 mL) were
charged to a 50 mL vessel in a glovebox. The reaction was stirred overnight and then
concentrated in vacuo to afford 14 as a solid which was used without further purification. In a
Uniqsis FlowSyn system, the lines and reactor plate were flushed with MeOH, and then feeds
1 and 2 were flushed with dry THF. Once the system was completely purged the lines were
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 23
1
2
3
4
5
6
filled with their respective feeds. 4-fluorobenzylzinc chloride (0.50 M in THF, 0.15 mL/min)
and 14 (0.25 M in THF, 0.30 mL/min) were fed into a Uniqsis glass chip (1.8 mL) at 20 °C.
The residence time was 4 minutes. The crude solution was then quenched into MeOH. Very
low conversion to product observed ca. 5%.
7
8
9
Synthesis of ketone 19. LiHMDS (1.00 M in THF) (577 mL, 577 mmol, 2.05 equiv) was
charged to a 1 L vessel which had been purged with nitrogen. The vessel was cooled to 10 °C
and a mixture of 2-chloro-4,6-dimethyl-pyridine (17) (40.0 g, 283 mmol, 1.00 equiv) and ethyl-
4-fluorobenzoate (18) (47.5 g, 283 mmol, 1.00 equiv) added dropwise using an addition funnel
over 1 h. Once complete, the addition funnel was rinsed with THF (20 mL) and the reaction
maintained at 10 °C overnight. The jacket temperature was set to ˗10 °C and the reaction
quenched by dropwise addition of aq. H2SO4 (2.00 M, 300 mL) over 1 h and the phases
separated. The aqueous phase was separated, and the batch was then distilled at 40 °C and 186
mbar over 1.5 h. Toluene (400 mL) was added and the mixture was distilled to ca. 400 mL at
55 °C and 160 mbar. Water (200 mL) was added and the layers separated. The aqueous was
removed and the organics heated to 60 °C before being filtered. The filtrate was recharged to
the jacketed vessel and distilled to ca. 200 mL at 60 °C and 100 mbar. The solution was
maintained at 60 °C and heptane (200 mL) added dropwise over 1 h. The solution was seeded
and stirring continued at 60 °C overnight. The slurry was cooled to 0 °C over 5 h and stirred
overnight at 0 °C. The slurry was filtered and the cake washed with heptane (120 mL) followed
by 2-propanol (120 mL). The cake was dried at 40 °C under vacuum to constant weight to give
ketone 19 as an off-white solid (62.6 g, 221 mmol, 78%). Assay (QNMR) 93% w/w.
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
A portion of the material was further purified by recrystallization from 2-propanol for
analytical data. 1H NMR (400 MHz, DMSO-d6): δ 8.15-8.09 (m, 2H), 7.44-7.35 (m, 2H), 7.24
13
(s, 1H), 7.16 (s, 1H), 4.48 (s, 2H), 2.43 (s, 3H); C NMR (101 MHz, DMSO-d6): δ 194.7,
165.2 (d, 1JC-F = 252.3 Hz), 158.6, 149.2, 148.3, 132.9 (d, 4JC-F = 2.8 Hz), 131.3 (d, 3JC-F = 9.6
2
19
Hz), 124.0, 122.6, 115.8 (d, JC-F = 22.0 Hz), 43.4, 23.4; F NMR (376 MHz, DMSO-d6): δ
˗105.43; HRMS: exact mass calculated for [M+H]+ (C14H12NOFCl) requires m/z 264.0591,
found m/z 264.0596.
Synthesis of chloro 20. Ketone 19 (200 mg, 0.758 mmol, 1.00 equiv), PTSA (mono-hydrate)
(7 mg, 0.0368 mmol, 0.05 equiv), and CH2Cl2 (1.5 mL) were charged to a flask and NCS (104
mg, 0.779 mmol, 1.03 equiv) added portionwise and the reaction stirred for 2 h. The reaction
was diluted with water (10 mL) and the organics extracted with CH2Cl2 (3 x 10 mL). The
organics were combined, dried over Na2SO4, filtered, and evaporated to afford a colourless oil.
The crude material was purified by flash silica column chromatography, eluent 15%
EtOAc/iso-hexane, to afford chloro 20 as a colourless oil (216 mg, 0.725 mmol, 96%).1H NMR
(500 MHz, CDCl3): δ 8.03-7.98 (m, 2H), 7.25 (s, 1H), 7.20-7.13 (m, 3H), 6.05 (s, 1H), 2.54 (s,
3H); 13C NMR (126 MHz, CDCl3): δ 188.9, 166.5 (d, 1JC-F = 258.0 Hz), 160.4, 151.3, 147.3,
132.2 (d, 3JC-F = 9.7 Hz), 130.0 (d, 4JC-F = 3.0 Hz), 121.3, 120.6, 116.5 (d, 2JC-F = 22.1 Hz), 58.3,
24.4; 19F NMR (470 MHz, CDCl3): δ -101.89; HRMS: exact mass calculated for [M+H]+
(C14H11NOFCl2) requires m/z 298.0202, found m/z 298.0204.
Synthesis of AZD4635 (1) (Scheme 4c). Chloro 20 (100 mg, 0.34 mmol, 1.00 equiv) and EtOH
(1 mL) were charged to a flask. 1-aminoguanidine bicarbonate (3) (46 mg, 0.34 mmol, 1.00
equiv) and NaI (10 mg, 0.07 mmol, 0.20 equiv) were added and the reaction stirred at room
temperature overnight. Analysis showed incomplete oxidation to the desired product and
therefore the reaction was heated to 75 °C for 2 h. The reaction was cooled to room temperature
and water (0.5 mL) added. The reaction was stirred for 2 h before being filtered under vacuum
until dry to afford AZD4635 (1) as a pale yellow solid (17 mg, 0.054 mmol, 16%, 1:23 = 9:1).
ACS Paragon Plus Environment

Page 17 of 23
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Synthesis of di-ketone 21 (Scheme 4d). Ketone 19 (100 mg, 0.379 mmol, 1.00 equiv),
Cu(OAc)2·H2O (4 mg, 0.0200 mmol, 5 mol%), K2CO3 (2 mg, 0.0145 mmol, 5 mol%) and
CH3CN (1 mL) were charged to a vial and the reaction stirred vigorously under air for 72 h.
77% 1H NMR assay yield, benzylbenzoate as standard. See optimized procedure (Scheme 6a)
for characterization data.
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 AZD4635 (1) (Scheme 4d). Diketone 21 (1.00g, 3.53 mmol, 1.00 equiv), 1-
aminoguanidine hydrochloride (22) (507 mg, 4.59 mmol, 1.30 equiv), and i-PrOH (20 mL)
were charged to a flask, TFA (1 mL, 13.1 mmol, 3.70 equiv) added and the reaction heated to
70 °C under air for 5 h. The reaction was allowed to cool to room temperature, quenched with
NEt3 (1.63 mL, 11.7 mmol, 3.32 equiv) and diluted with water (30 mL). The reaction was
stirred for 16 h and filtered to afford AZD4635 (1) as an orange solid (403 mg, 1.28 mmol,
1:23 = 4:1, Assay (QNMR) 89.5%, 32% corrected yield).
Synthesis of iodo 24. BPin 7 (200 mg, 0.625 mmol, 1.00 equiv) was charged to a 3-neck flask
fitted with reflux condenser and the flask purged with nitrogen (x 3). THF (2 mL) and water (2
mL) were added and the flask wrapped in aluminium foil. NaI (468 mg, 3.12 mmol, 5.00 equiv)
was added, followed by chloramine T (527 mg, 1.87 mmol, 3.00 equiv) and the reaction
brought to 80 °C for 4 h. The reaction was allowed to cool to room temperature before being
quenched by dropwise addition of Na2S2O3·5(H2O) (620 mg, 2.50 mmol, 4.00 equiv) in water
(8 mL). The organics were extracted with EtOAc (3 x 20 mL), combined, dried over Na2SO4,
filtered, and evaporated to afford a beige solid. The crude material was purified by flash silica
column chromatography, eluent EtOAc/iso-hexane, to afford iodo 24 as a white solid (107 mg,
0.422 mmol, 68%). 1H NMR (400 MHz, CDCl3): δ 7.54 (s, 1H), 7.47 (s, 1H), 2.48 (s, 3H); 13C
NMR (101 MHz, CDCl3): δ 160.0, 150.8, 131.1, 129.9, 106.9, 23.9; HRMS: exact mass
calculated for [M+H]+ (C6H6NClI) requires m/z 253.9233, found m/z 253.9235.
Synthesis of alkyne 26. Iodo 24 (394 mg, 1.55 mmol, 1.00 equiv), 1-ethynyl-4-fluoro-benzene
(25) (206 mg, 1.72 mmol, 1.10 equiv), Pd(PPh3)2Cl2 (55 mg, 0.0784 mmol, 5 mol%), and CuI
(20 mg, 0.105 mmol, 6 mol%) were charged to a flask and the flask purged with nitrogen (x
3). 1,4-dioxane (8 mL) was added, followed by NEt3 (8 mL) and the reaction placed in a pre-
heated block at 80 °C for 20 minutes. The reaction was allowed to cool to room temperature,
diluted with water (20 mL) and extracted with EtOAc (3 x 20 mL). The organics were
combined, dried over Na2SO4, filtered, and evaporated to afford a pale brown oil. The crude
material was purified by flash silica column chromatography, eluent 0-5% EtOAc/iso-hexane,
1
to afford alkyne 26 as a pale yellow solid (345 mg, 1.40 mmol, 90%). H NMR (400 MHz,
CDCl3): δ 7.55-7.48 (m, 2H), 7.23 (s, 1H), 7.15 (s, 1H), 7.03-7.12 (m, 2H), 2.53 (s, 3H); 13C
NMR (101 MHz, CDCl3): δ 163.3 (d, 1JC-F = 251.6 Hz), 159.5, 150.9, 134.3, 134.1 (d, 3JC-F
=
8.6 Hz), 123.7, 122.9, 118.0 (d, 4JC-F = 3.6 Hz), 116.1 (d, 2JC-F = 22.3 Hz), 93.7, 85.7 (d, 5JC-F
19
= 1.4 Hz), 24.2; F NMR (376 MHz, CDCl3): δ -108.69; HRMS: exact mass calculated for
[M+H]+ (C14H10NClF) requires m/z 246.0486, found m/z 246.0490.
Synthesis of AZD4635 (1) (Scheme 4e). Alkyne 26 (54 mg, 0.220 mmol, 1.00 equiv) and DMSO
(2.2 mL) were charged to a microwave tube. I2 (11 mg, 0.0433 mmol, 0.20 equiv) was added,
the tube sealed with a suba seal, and the reaction brought to 130 °C in a heating block for 34 h.
The reaction was allowed to cool to room temperature and 1-aminoguanidine bicarbonate (3)
(30 mg, 0.220 mmol, 1.00 equiv) added. The reaction was stirred at room temperature for 2 h,
then brought to 80 °C for a further 6 h. The reaction was allowed to cool to room temperature,
diluted with water (30 mL) and the organics extracted with EtOAc (3 x 20 mL). The organics
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 23
1
2
3
4
5
6
were combined, dried over Na2SO4, filtered, and evaporated to afford an orange residue. The
residue was taken up in iso-hexane (10 mL) and filtered. The residue was then taken up in
toluene, heated gently with a heat gun and filtered. The filtrate was concentrated in vacuo to
afford AZD4635 (1) as an orange gum (35 mg, 0.111 mmol, 50%, 1:23 = 3:1 by 1H NMR).
7
8
9
Synthesis of bromo 29. 1,2,4-triazin-3-amine (27) (150 mg, 1.56 mmol, 1.00 equiv), TBACl
(434 mg, 1.56 mmol, 1.00 equiv), and K2CO3 (432 mg, 3.13 mmol, 2.00 equiv) were charged
to a flask and the flask purged with nitrogen (x 4). MeCN (1.6 mL) was added, followed by
dropwise addition of a solution of NBS (333 mg, 1.87 mmol, 1.20 equiv) in MeCN (3 mL).
The reaction was stirred for 10 minutes before being quenched with a solution of Na2S2O3 (155
mg, 0.625 mmol, 0.40 equiv) in water (1 mL) and stirred for 10 minutes. The reaction was
diluted with water (10 mL) and EtOAc (20 mL). The layers were separated and the organics
washed with 2 M HCl (2 x 10 mL). The acidic washes were combined and basified with 2 M
NaOH before being extracted with EtOAc (3 x 20 mL). The organics were then combined,
dried over Na2SO4, filtered, and evaporated to afford a pale purple solid. The crude material
was purified by flash silica column chromatography, eluent 40% EtOAc/CH2Cl2, to afford
bromo 29 as a beige solid (218 mg, 1.25 mmol, 80%). 1H NMR (500 MHz, DMSO-d6): δ 8.40
(s, 1H), 7.47 (br s, 2H); 13C NMR (126 MHz, DMSO-d6): δ 162.3, 152.6, 135.5; HRMS: exact
mass calculated for [M+H]+ (C3H4N4Br) requires m/z 174.9619, found m/z 174.9618. The
spectroscopic data are in agreement with the reported literature.27
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
Competitive Suzuki Miyaura experiment. Bromo 29 (38 mg, 0.217 mmol, 1.00 equiv), BPin 7
(70 mg, 0.219 mmol, 79% w/w, 1.00 equiv), Pd(OAc)2 (2 mg, 0.00891 mmol, 4 mol%), APhos
(6 mg, 0.0226 mmol, 10 mol%), and K2CO3 (45 mg, 0.326 mmol, 1.50 equiv) were charged to
a microwave tube and the tube sealed with a suba seal and purged with nitrogen (x 3). 1,4-
Dioxane (0.6 mL) was added, followed by 1-bromo-4-fluorobenzene 30 (24 µL, 0.220 mmol,
1.00 equiv), and water (0.2 mL). The reaction was placed in a pre-heated block at 80 °C for 4
h. The reaction was allowed to cool to room temperature, filtered through celite eluting with
EtOAc, and concentrated in vacuo. Triazine coupling product 31: 60%, 1-bromo-4-
fluorobenzene coupling product 32: 9%. Yield by QNMR (1,4-dinitrobenzene as internal
standard).
Analytical samples obtained by silica column chromatography (eluent EtOAc/iso-hexane):
31: 1H NMR (500 MHz, DMSO-d6): δ 8.92 (s, 1H), 7.90 (s, 1H), 7.87 (s, 1H), 2.51 (s, 3H); 13C
NMR (126 MHz, DMSO-d6): δ 162.8, 159.8, 150.3, 148.9, 145.4, 144.4, 117.4, 116.1, 23.7;
HRMS: exact mass calculated for [M+H]+ (C9H9N5Cl) requires m/z 222.0546, found m/z 222.0539.
32: 1H NMR (400 MHz, DMSO-d6): δ 7.94-7.86 (m, 2H), 7.63 (s, 1H), 7.62 (s, 1H), 7.39-7.32
(m, 2H), 2.51 (s, 3H); 13C NMR (101 MHz, DMSO-d6): δ 163.1 (d, 1JC-F = 247.5 Hz), 159.6,
150.3, 149.8, 132.4 (d, 4JC-F = 3.1 Hz), 129.4 (d, 3JC-F = 8.6 Hz), 119.9, 118.3, 116.1 (d, 2JC-F
=
19
21.7 Hz), 23.6; F NMR (376 MHz, DMSO-d6): δ -111.87; HRMS: exact mass calculated for
[M+H]+ (C12H10NFCl) requires m/z 222.0486, found m/z 222.0481.
Synthesis of bromo 33. Bromo 29 (96 mg, 0.549 mmol, 1.00 equiv) and acetic acid (1.5 mL)
were charged to a vial and H2O2 (33% w/w in H2O) (0.1 mL, 1.10 mmol, 2.00 equiv) added
dropwise. The reaction was stirred for 24 h, before being diluted with iso-hexane (10 mL) and
filtered. The filtercake was washed with further iso-hexane (2 x 20 mL) before being taken up
in MeOH (20 mL) and the solvent removed in vacuo to afford 33 as a white solid (66 mg, 0.346
mmol, 63%). 1H NMR (500 MHz, DMSO-d6): δ 12.51 (s, 1H, OH), 7.07 (br s, 2H, NH2); 13C
NMR (126 MHz, DMSO-d6): δ 158.5, 155.9, 135.6; HRMS: exact mass calculated for [M+H]+
ACS Paragon Plus Environment

Page 19 of 23
Organic Process Research & Development
1
2
3
4
(C3H4N4OBr) requires m/z 190.9568, found m/z 190.9572.
5
6
7
8
Procedure for mono-ketone screening (Table 1). Reactions were performed in oven dried tubes
at 20 °C under nitrogen. Pyridine 17 (1.00 g, 7.06 mmol, 1.00 equiv), 4-fluorobenzoate (18)
(1.19 g, 7.08 mmol, 1.00 equiv), and the required solvent (7 mL) were charged to a tube and
the required base added over ca. 5 minutes at room temperature. The reactions were stirred for
4 h before being analysed by HPLC.
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 di-ketone 21 (Scheme 6a). Ketone 19 (200 mg, 0.758 mmol, 1.00 equiv), Cu(OAc)2
(14 mg, 0.0771 mmol, 10 mol%), K2CO3 (4 mg, 0.0404 mmol, 5 mol%), and MeCN (2.5 mL)
were charged to a vial. TMEDA (6 µL, 0.0400 mmol, 5 mol%) was added and air bubbled
through the reaction for 20 h. The reaction was diluted with water (10 mL) and extracted with
EtOAc (3 x 10 mL). The organics were combined, dried over Na2SO4, filtered, and evaporated
to afford a green oil which solidified on standing. The crude material was purified by flash
silica column chromatography, eluent 15% EtOAc/iso-hexane, to afford diketone 21 as a pale
1
yellow solid (168 mg, 0.605 mmol, 80%). H NMR (400 MHz, DMSO-d6): δ 8.14-8.08 (m,
13
2H), 7.73 (s, 2H), 7.51-7.43 (m, 2H), 2.56 (s, 3H); C NMR (126 MHz, DMSO-d6) 190.8,
1
3
189.9, 166.3 (d, JC-F = 255.7 Hz), 161.2, 150.3, 142.2, 133.6 (d, JC-F = 10.2 Hz), 128.7 (d,
4JC-F = 2.6 Hz), 121.6, 120.3, 116.5 (d, 2JC-F = 22.4 Hz) , 23.62; 19F NMR (470 MHz, DMSO-
d6): δ ˗101.50; HRMS: exact mass calculated for [M+H]+ (C14H10NO2FCl) requires m/z
278.0384, found m/z 278.0383.
Synthesis of 38. 2-Chloro-4,6-dimethylpyridine (17) (200 mg, 1.41 mmol, 1.00 equiv) and ethyl
4-fluorobenzoate (18) (238 mg, 1.42 mmol, 1.00 equiv) were charged to a flask which had been
flushed with nitrogen, cooled to 0 °C, and LiHMDS (1 M in THF) (3.2 mL, 3.20 mmol, 2.30
equiv) added dropwise. After 1 h a solution of 37 (358 mg, 1.55 mmol, 1.10 equiv) in THF (2.8
mL) was added dropwise. The reaction was stirred at room temperature for 15 h before being
diluted with water (30 mL) and EtOAc (30 mL). The layers were separated and the aqueous
extracted with further EtOAc (2 x 30 mL). The organics were combined, dried over Na2SO4,
filtered, and evaporated to afford a pale orange oil. The crude material was purified by flash
silica column chromatography, eluent 20% EtOAc/iso-hexane to afford 38 as a white solid (367
1
mg, 0.743 mmol, 53%). H NMR (500 MHz, DMSO-d6, 90 °C): δ 8.90 (br s, 1H), 8.17-8.04
13
(m, 2H), 7.36-7.24 (m, 4H), 6.53 (s, 1H), 2.24 (s, 3H), 1.34 (s, 9H), 1.26 (s, 9H); C NMR
1
(126 MHz, DMSO-d6, 90 °C): δ 191.5, 164.8 (d, JC-F = 252.8 Hz), 158.2, 154.3 (br), 148.8,
131.2 (d, 3JC-F = 9.5 Hz), 123.0, 121.8, 115.2 (d, 2JC-F = 22.0 Hz), 81.2, 79.2, 65.3 (br), 27.3,
27.3, 23.1; 19F NMR (470 MHz, DMSO-d6, 90 °C): δ ˗105.46; HRMS: exact mass calculated
for [M+H]+ (C24H30N3O5FCl) requires m/z 494.1858, found m/z 494.1863.
Synthesis of di-ketone 21 (Scheme 6b). 38 (55 mg, 0.111 mmol, 1.00 equiv) and CH2Cl2 (1 mL)
were charged to a flask and TFA (0.1 mL, 1.00 M) added dropwise. After 24 h the reaction was
diluted with water (10 mL) and extracted with CH2Cl2 (3 x 20 mL). The organics were
combined, dried over Na2SO4, filtered, and evaporated to afford a pale yellow residue. The
crude material was purified by flash silica column chromatography, eluent 15% EtOAc/iso-
hexane, to afford di-ketone 21 as an off-white solid (26 mg, 0.0936 mmol, 84%).
Flow synthesis of 38. In a VAPOURTEC E-series, the lines were flushed with THF before a
mixture of 17 (18 mL, 1.00 M in THF, 18.0 mmol) and 18 (18 mL, 1.00 M in THF, 18.0 mmol)
(0.30 mL/min), and LiHMDS (42 mL, 1.00 M in THF, 42.0 mmol, 0.70 mL/min) were fed into
a reactor 1 (standard coiled tubular reactor, 10 mL) at 40 °C. The residence time was 10
minutes. This solution and a solution of 37 (51 mL, 0.50 M in THF, 25.5 mmol, 0.85 mL/min)
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 23
1
2
3
4
5
6
7
8
9
were then fed into reactor 2 (glass microreactor, 0.6 mL) at – 40 °C and the residence time was
20 seconds. The crude solution was then quenched into AcOH. The solution was diluted with
MTBE and water/brine and the layers separated. The organics were concentrated in vacuo and
the crude material recrystallized from acetone/water to afford 38 as a white solid (7.66 g, 15.5
mmol, Assay (QNMR) 73% w/w, 63% corrected 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
Synthesis of AZD4635 (1) (Scheme 6d). Di-ketone 21 (50 mg, 0.180 mmol, 1.00 equiv) and
IPA (0.5 mL) were charged to a tube. Zn(OTf)2 (5 mg, 0.0233 mmol, 13 mol%) was added,
followed by aminoguanidine hydrochloride (22) (20 mg, 0.181 mmol, 1.00 equiv) and the
reaction brought to 50 °C for 24 h. HPLC analysis showed regioisomeric ratio 94:6 (1:23) with
ca. 50% conversion.
Process for purging of regioisomer 23. AZD4635 (1.06 g, 3.36 mmol, 1:23 = 98:2) was slurried
in PhMe (10 mL) at 100 °C, allowed to cool to room temperature and stirred for 10 minutes.
The material was then filtered and the cake washed with PhMe (10 mL). The cake was slurried
in further PhMe (10 mL), filtered, and the cake washed with PhMe (2 x 10 mL) and filtered
under vacuum until dry to afford AZD4635 as an off-white solid (770 mg, 2.44 mmol, 73%,
1:23 = 99.96:0.04).
Procedure for halogenation screening (Scheme 7). Reactions were performed in 4 mL vials in
a 24-well plate format, situated in an inerted glovebox (< 10 ppm O2 and < 1 ppm H2O). 4 (50
mg, 0. mmol, 1.00 equiv) and all other solid reagents were dispensed into all vials via a Quantos
QX96 weighing robot (accuracy +/- 2%). Stir disks were added to all vials. Degassed solvent
(1 mL) was added to each vial, followed by addition of liquid reagents. The reactions were
sealed and heated to 25 °C for 2 h and then 65 °C for a further 18 h. Conversion by HPLC. See
optimized procedure (Table 2) for characterization data.
Procedure for bromination screening (Table 2). Reactions were performed in 1 mL vials in a
96-well plate format, situated in an inerted glovebox (< 10 ppm O2 and < 1 ppm H2O). From a
stock solution of m-terphenyl (0.04 M in toluene) a known volume was dispensed into all vials
before evaporation of the solvent. 4 (20 mg, 0.105 mmol, 1.00 equiv) was dispensed into all
vials via a Quantos QX96 weighing robot (accuracy +/- 0.5 mg), followed by TBACl and
K2CO3. Stir disks were added to all vials. Degassed solvent (250 µL) was added to each vial,
followed by addition of the desired base. The reactions were stirred for 10 minutes before
addition of the desired brominating agent in a solution of MeCN (150 µL) via a multi-dispense
pipette. The reactions were sealed and heated to 25 °C for 20 h.
Synthesis of 5 (Scale-up of screening hit for isolation, Table 2). Triazine 4 (200 mg, 1.05 mmol,
1.00 equiv) and K2CO3 (291 mg, 2.11 mmol, 2.00 equiv) were charged to a flask and the flask
sealed with a suba seal and purged with nitrogen (x 3). MeCN (1 mL) was added, followed by
dropwise addition of a solution of NBS (225 mg, 1.26 mmol, 1.20 equiv) in MeCN (2 mL) at
room temperature and the reaction stirred for 0.5 h. The reaction was quenched with a solution
of Na2S2O3·5H2O (261 mg, 1.05 mmol, 1.00 equiv) in water (4 mL) and stirring continued for
0.5 h. The reaction was diluted with EtOAc (10 mL) and the layers separated. The aqueous was
further extracted with EtOAc (2 x 10 mL), the organics combined, dried over Na2SO4, filtered,
and evaporated to afford a brown solid. The crude material was purified by flash silica column
chromatography, eluent 10-30% EtOAc/iso-hexane to afford bromo 5 as a beige solid (180 mg,
0.669 mmol, 64%). 1H NMR (400 MHz, DMSO-d6): δ 7.86-7.80 (m, 2H), 7.54 (br s, 2H), 7.40-
13
1
7.34 (m, 2H); C NMR (101 MHz, DMSO-d6): δ 163.4 (d, JC-F = 248.8 Hz), 162.3, 157.3,
134.5, 131.8 (d, 3JC-F = 8.9 Hz), 131.5 (d, 4JC-F = 3.1 Hz), 115.3 (d, 2JC-F = 22.0 Hz); 19F NMR
ACS Paragon Plus Environment

Page 21 of 23
Organic Process Research & Development
1
2
3
4
5
(376 MHz, DMSO-d6): δ ˗109.64; HRMS: exact mass calculated for [M+H]+ (C9H7N4FBr)
requires m/z 268.9838, found m/z 268.9840.
6
7
Associated Content
8
9
Supporting Information
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
Tables of optimization data and NMR spectra of all compounds.
Author Information
Corresponding Author
E-mail: mairi.littleson1@astrazeneca.com
ORCID
Mairi M. Littleson:
Staffan Karlsson:
David J. Klauber:
Thomas O. Ronson:
Acknowledgements
We thank Andrew Dalgleish and Edward Deane for contributions to early work, Andrew R.
Turner for NMR support, Alexander Abbott for HRMS support, and James Douglas for helpful
discussion.
References
(1) Cornwall, P.; Diorazio, L. J.; Monks, N. Route design, the foundation of successful
chemical development. Bioorg. Med. Chem. 2018, 26, 4336.
(2) Fredholm, B. B.; Ijzerman, A. P.; Jacobson, K. A.; Klotz, K-N.; Linden, J. International
Union of Pharmacology. XXV. Nomenclature and Classification of Adenosine Receptors.
Pharmacol. Rev. 2001, 53, 527.
(3) (a) Congreve, M.; Andrews, S. P.; Doré, A. S.; Hollenstein, K.; Hurrell, E.; Langmead, C.
J.; Mason J. S.; Ng, I. W.; Tehan, B.; Zhukov, A.; Weir, M.; Marshall, F. H. Discovery of
1,2,4-Triazine Derivatives as Adenosine A2A Antagonists using Structure Based Drug
Design. J. Med. Chem. 2012, 55, 1898. (b) Langmead, C. J.; Andrews, S. P.; Congreve,
M.; Errey, J. C.; Hurrell, E.; Marshall, F. H.; Mason, J. S.; Richardson, C. M.; Robertson,
N.; Zhukov, A.; Weir, M. Identification of Novel Adenosine A2A Receptor Antagonists by
Virtual Screening. J. Med. Chem. 2012, 55, 1904.
(4) Congreve, M. S.; Philippe, S.; Jonathan, A.; Mason, S.; Richardson, C. M.; Brown, G. A.;
1,2,4-triazine-4-amine derivatives. WO2011/95625 A1, 2011.
(5) Douglas, J. J.; Adams, B. W. V.; Benson, H.; Broberg, K.; Gillespie, P. M.; Hoult, O.;
Ibraheem, A. K.; Janbon, S.; Janin, G.; Parsons, C. D.; Sigerson, R. C.; Klauber, D. J.
Multi-kilo scale preparation of AZD4635 via C-H borylation and bromination: The
corrosion of Tantalum by a bromine/methanol mixture. Org. Process Res. Dev. 2019, 23,
62.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 23
1
2
3
4
5
6
(6) Cascioferro, S.; Parrino, B.; Spanò, V.; Carbone, A.; Montalbano, A.; Barraja, P.; Diana,
P.; Cirrincione, G. An overview on the recent developments of 1,2,4-triazine derivatives
as anticancer compounds. Eur. J. Med. Chem. 2017, 142, 328.
(7) Bøgevig, A.; Federsel, H-J.; Huerta, F.; Hutchings, M. G.; Kraut, H.; Langer, T.; Löw, P.;
Oppawsky, C.; Rein, T.; Saller, H. Route Design in the 21st Century: The ICSYNTH
Software Tool as an Idea Generator for Synthesis Prediction. Org. Process Res. Dev. 2015,
19, 357.
(8) Raw, S. A.; Taylor, R. J. K. Advances in Heterocyclic Chemistry, vol 100. Elsevier Press,
2010. p. 75.
(9) Murakami, K.; Yamada, S.; Kaneda, T.; Itami, K. C-H Functionalisation of Azines. Chem.
Rev. 2017, 117, 9302.
(10) See ESI for details.
(11) Hilton, M. C.; Dolewski, R. D.; McNally, A. Selective Functionalisation of Pyridines
via Heterocyclic Phosphonium Salts. J. Am. Chem. Soc. 2016, 138, 13806.
(12) A recent publication from the McNally group reports that 2,6-disubstituted pyridines
are problematic substrates in this reaction. Hilton, M. C.; Zhang, X.; Boyle, B. T.; Alegre-
Requena, J. V.; Paton, R. S.; McNally, A. Heterobiaryl synthesis by contractive C-C
coupling via P(V) intermediates. Science 2018, 362, 799.
(13) Biftu, T.; Feng, D.; Ponpipom, M.; Girotra, N.; Liang, G.-B.; Qian, X.; Bugianesi, R.;
Simeone, J.; Chang, L.; Gurnett, A.; Liberator, P.; Dulski, P.; Leavitt, P. S.; Crumley, T.;
Misura, A.; Murphy, T.; Rattray, S.; Samaras, S.; Tamas, T.; Mathew, J.; Brown, C.;
Thompson, D.; Schmatz, D.; Fisher, M.; Wyvratt, M. Synthesis and SAR of 2,3-
diarylpyrrole inhibitors of parasite cGMP-dependent protein kinase as novel anticoccidial
agents. Bioorg. Med. Chem. Lett. 2005, 15, 3296.
(14) For selected papers, see: (a) Chen, C.-Y.; Hu, M.-C.; Liu, M.-C.; Yan, P.-C.; Wang, J.-
J.; Chung, M.-I. Efficient synthesis of quinoxalines with hypervalent iodine as a catalyst.
Tetrahedron 2013, 69, 9735–9741. (b) Viswanadham, K. K. D. R.; Reddy, M. P.;
Sathyanarayana, P.; Ravi, O.; Kant, R.; Bathula, S. R. Iodine-mediated oxidative
annulation for one-pot synthesis of pyrazines and quinoxalines using a multipathway
coupled domino strategy. Chem. Commun. 2014, 50, 13517–13520. (c) Chan, C.-K.;
Chang, M.-Y. N-Bromosuccinimide-Mediated Synthesis of Substituted Quinoxalines.
Synthesis 2016, 48, 3785.
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
(15) Battagliarin, G.; Li. C.; Enkelmann, V.; Müllen, K. 2,5,8,11-Tetraboronic Ester
Perylenediimides: A Next Generation Building Block for Dye-Stuff Synthesis. Org. Lett.
2011, 13, 3012.
(16) For selected papers, see: (a) Motti, E.; Mignozzi, A.; Catellani, M. A new type of
palladium-catalysed aromatic cross-coupling combined with a Suzuki reaction: synthesis
of selectively 2,3’substituted 1,1’;2’,1’’-terphenyl derivatives. J. Mol. Catal. A: Chem.
2003, 204–205, 115. (b) Motti, E.; Della Ca’, N.; Deledda, S.; Fava, E.; Panciroli, F.;
Catellani, M. Palladium-catalyzed unsymmetrical aryl couplings in sequence leading to o-
teraryls: dramatic olefin effect on selectivity. Chem. Commun. 2010, 46, 4291. (c) Della
Ca’, N.; Fontana, M.; Motti, E.; Catellani, M. Pd/Norbornene: A Winning Combination for
Selective Aromatic Functionalization via C–H Bond Activation. Acc. Chem. Res. 2016, 49,
1389.
(17) For selected papers, see: (a) Maestri, G.; Motti, E.; Della Ca’, N.; Malacria, M.; Derat,
E.; Catellani, M. Of the Ortho Effect in Palladium/Norbornene-Catalysed Reactions: A
Theoretical Investigation. J. Am. Chem. Soc. 2011, 133, 8574. (b) Wang, J.; Li, R.; Dong,
Z.; Liu, P.; Dong, G. Complementary site-selectivity in arene functionalization enabled by
overcoming the ortho constraint in palladium/norbornene catalysis. Nat. Chem. 2018, 10,
866.
ACS Paragon Plus Environment

Page 23 of 23
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
(18) Dong, Z.; Lu, G.; Wang, J.; Liu, P.; Dong, G. Modular ipso/ortho Difunctionalization
of Aryl Bromides via Palladium/Norbornene Cooperative Catalysis. J. Am. Chem. Soc.
2018, 140, 8551.
(19) Blanchot, M.; Candito, D. A.; Larnaud, F.; Lautens, M. Formal Synthesis of Nitidine
and NK109 via Palladium-Catalyzed Domino Direct Arylation/N-Arylation of Aryl
Triflates. Org. Lett. 2011, 13, 1486.
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
(20) Dudfield, P. J.; Le, V.-D.; Lindell, S. D.; Rees, C. W. Synthesis of C-ribosyl 1,2,4-
triazolo[3,4-f][1,2,4]triazines as inhibitors of adenosine and AMP deaminases. J. Chem.
Soc., Perkin Trans. 1, 1999, 2937.
(21) Zhai, D.-D.; Zhang, X.-Y.; Liu, Y.-F; Zheng, L.; Guan, B.-T. Potassium Amide-
Catalyzed Benzylic C-H Bond Addition of Alkylpyridines to Styrenes. Angew. Chem. Int.
Ed. 2018, 57, 1650.
(22) Osterberg, P. M.; Niemeier, J. K.; Welch, C. J.; Hawkins, J. M.; Martinelli, J. R.;
Johnson, T. E.; Root, T. W.; Stahl, S. S. Experimental Limiting Oxygen Concentrations
for Nine Organic Solvents at Temperatures and Pressures Relevant to Aerobic Oxidations
in the Pharmaceutical Industry. Org. Process Res. Dev. 2015, 19, 1537.
(23) Steves, J. E.; Preger, Y.; Martinelli, J. R.; Welch, C. J.; Root, T. W.; Hawkins, J. M.;
Stahl, S. S. Process Development of CuI/ABNO/NMI-Catalysed Aerobic Alcohol
Oxidation. Org. Process Res. Dev. 2015, 19, 1548.
(24) Sivakamasundari, S.; Ganesan, R. Kinetics and Mechanism of the Bromination of
Aromatic Compounds by N-Bromosuccinimide in Solution. Int. J. Chem. Kinet. 1980, 12,
837.
(25) Butters, M.; Catterick, D.; Craig, A.; Curzons, A.; Dale, D.; Gillmore, A.; Green, S. P.;
Marziano, I.; Sherlock, J-P.; White, W. Critical Assessment of Pharmaceutical Processes–
A Rationale for Changing the Synthetic Route. Chem. Rev. 2006, 106, 3002.
(26) Leng, R. B.; Emonds, M. V. M.; Hamilton, C. T.; Ringer, J. W. Holistic Route
Selection. Org. Process Res. Dev. 2012, 16, 415.
(27) Prescher, J. A.; Houk, K. N.; Mehl, R. A.; Liu, F.; Blizzard, R. J.; Liang, Y.; Kamber,
D. N. 1,2,4-Triazines Are Versatile Bioorthogonal Reagents. J. Am. Chem. Soc. 2015, 137,
8388.
ACS Paragon Plus Environment