Development of a Convergent Large-Scale Synthesis for Venetoclax, a First-in-Class BCL-2 Selective Inhibitor

Article abstract of DOI:10.1021/acs.joc.8b02750

The process development of a new synthetic route leading to an efficient and robust synthetic process for venetoclax (1: the active pharmaceutical ingredient (API) in Venclexta) is described. The redesigned synthesis features a Buchwald-Hartwig amination to construct the core ester 23c in a convergent fashion by connecting two key building blocks (4c and 26), which is then followed by a uniquely effective saponification reaction of 23c using anhydrous hydroxide generated in situ to obtain 2. Finally, the coupling of the penultimate core acid 2 with sulfonamide 3 furnishes drug substance 1 with consistently high quality. The challenges and solutions for the key Pd-catalyzed C-N cross-coupling will also be discussed in detail. The improved synthesis overcomes many of the initial scale-up challenges and was accomplished in 46% overall yield from 3,3-dimethyldicyclohexanone (6), more than doubling the overall yield of the first generation route. The new process was successfully implemented for producing large quantities of 1 with >99% area purity.

Full text of DOI:10.1021/acs.joc.8b02750

Subscriber access provided by Iowa State University | Library
Article
Development of a Convergent Large-Scale Synthesis for
Venetoclax, a First-in-Class BCL-2 Selective Inhibitor
Yi-Yin Ku, Vincent S Chan, Alan Christesen, Timothy Grieme,
Mathew Mulhern, Yu-Ming Pu, and Michael D. Wendt
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.8b02750 • Publication Date (Web): 07 Jan 2019
Downloaded from http://pubs.acs.org on January 8, 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 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Development of a Convergent Large-Scale Synthesis
for Venetoclax, a First-in-Class BCL-2 Selective
Inhibitor
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
Yi-Yin Ku,1* Vincent S. Chan, 3 Alan Christesen, 1 Timothy Grieme, 1 Mathew Mulhern, 1 Yu-
Ming Pu1 and Michael D Wendt 2
1 Process Chemistry, Pharmaceutical Research and Development, Abbvie.
2. Medicinal Chemistry, Pharmaceutical Research and Development, Abbvie
3. Current contact information: vchan@seagen.com
1 North Waukegan Road
North Chicago, IL 60064, United States
Venclexta™, Venetoclax, BCL-2 Inhibitor, ABT-199, Manufacture Process
NO2
NO2
H
H
N
N
NO2
O
Cl
O
O
S
O
O
S
NH2
O
S
O
NH
O
O
3
NH2
O
24b
CO2H
N
N
O
H
H
N
N
N
2HCl
N
N
H
N
CO2tBu
O
N
N
N
N
H
Br
Cl
4c
Cl
1. ABT-199
Venetoclax
26
Cl
2
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 49
1
2
3
4
5
6
7
8
9
ABSTRACT: The process development of a new synthetic route leading to an efficient and robust
synthetic process for Venetoclax (1: the active pharmaceutical ingredient (API) in Venclexta™)
is described. The redesigned synthesis features a Buchwald-Hartwig amination to construct the
core ester 23c in a convergent fashion by connecting two key building blocks (4c and 26), which
is then followed by a uniquely effective saponification reaction of 23c using anhydrous
hydroxide generated in-situ to obtain 2. Finally, the coupling of the penultimate core acid 2 with
sulfonamide 3 furnishes drug substance 1 with consistently high quality. The challenges and
solutions for the key Pd-catalyzed C-N cross-coupling will also be discussed in detail. The
improved synthesis overcomes many of the initial scale-up challenges and was accomplished in
46% overall yield from 3,3-dimethyldicyclohexanone 6, more than doubling the overall yield of
the first generation route. The new process was successfully implemented for producing large
quantities of 1 with >99% area purity.
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. INTRODUCTION
Venetoclax 1 (Figure 1), 1-4 the active pharmaceutical ingredient (API) of Venclexta™, obtained
accelerated approval from the U.S. Food and Drug Administration (FDA) ) in April 2016 1 as
the first small molecule BCL-2 selective inhibitor for the treatment of relapsed/refractory chronic
lymphocytic leukemia (CLL) patients with the 17p deletion genetic mutation.Venclexta is
currently being investigated in a number of other indications. Recently, it obtained FDA
approval under priority review as a combination therapy with rituximab for treating patients with
CLL or small lymphocytic lymphoma (SLL) with or without the 17p deletion, who have received
at least one prior therapy.1
2
ACS Paragon Plus Environment

Page 3 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
More than 30 years ago, BCL-2 proteins were discovered to be master regulators of apoptosis,2,5-
6 the highly regulated process by which the human body disposes damaged and unwanted cells.7-
8 The evasion of apoptosis is a hallmark of cancer,9 and its impairment is a key step in tumor
initiation and progression. BCL-2 proteins are antagonists of pro-apoptotic proteins and are
frequently overexpressed in cancers, and are implicated in driving both tumor growth as well as
resistance to chemotherapies.10-13 As a therapeutic target, it stood to reason that small molecule
inhibition of BCL-2 should lead to disruption of those detrimental protein-protein interactions
and restore apoptotic function. Herculean and innovative efforts by the medicinal chemistry team
at AbbVie (previously known as Abbott) spanning more than twenty years of research14-17
ultimately demonstrated the proof-of-concept that BCL-2 was a viable target for treating heme
malignancies which subsequently led to the discovery of 1, as a BCL-2 selective inhibitor.
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
NO2
H
N
O
O S
O
NH
O
O
N
N
H
N
N
1. ABT-199
Venetoclax
Cl
Figure 1. Venetoclax 1 (ABT-199)
A first generation synthesis adapted from the medicinal chemistry route was rapidly developed to
produce multi-kilogram quantities of 1. However, this route still was inefficient and involved
significant operational challenges as well as low overall yield and throughput. By strategically
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 49
1
2
3
4
5
6
7
8
9
redesigning the synthesis and implementing a Buchwald-Hartwig amination, the new synthetic
route successfully addressed the chemistry challenges associated with the first generation route
and improved the process convergence, overall yield and manufacturing robustness. This article
will discuss an enhanced, robust process for synthesizing 1 with consistently high quality..
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. FIRST GENERATION LARGE-SCALE SYNTHESIS
In general, developing an optimized chemical process for a drug candidate requires substantial
amounts of time and resources, which may not be appropriate for the early stages of development
as the physical supply of API is often on the critical path. Expeditious delivery of sufficient
quantities of API to enable advancement of a program oftentimes takes priority, especially
considering the high attrition rate in early development prior to demonstrating proof-of-concept
and intended efficacy.18 With this in mind, an initial assessment of the synthetic route indicated
that the quickest path to the development of a scalable process for the complex drug candidate 1
was to improve upon the existing medicinal chemistry approach (Scheme 1).19
Scheme 1. Medicinal Chemistry Approach to 1
NO2
O
OMe
F
NO2
H
H
N
CO2Me
O
N
O
O
O
S
O
S
N
NH2
N
O
NH
O
H
O
3
F
15
F
4a
O
HO
CO2H
H
N
N
O
N
H
N
N
N
N
TIPS
16
N
N
N
H
H
21
N
O
N
Cl
1. ABT-199
Venetoclax
Cl
Cl
5
2
4
ACS Paragon Plus Environment

Page 5 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
In the medicinal chemistry synthesis, the final coupling of the core acid 2 and the sulfonamide 3
worked reasonably well. However, the SNAr reaction for the key building block 4a had regio and
chemoselectivity concerns, and it also suffered from reproducibility issues caused by strong
scale-dependence of agitation; the synthesis of the key aldehyde intermediate 5 was lengthy and
included tedious purification. Additionally, the aldehyde 5 was a thick syrup, which made the
isolation of acceptable quality material difficult for large-scale production. Early process
chemistry efforts were therefore focused on improving these problematic and inefficient
reactions and isolations, and to overcome scale-up challenges associated with the medicinal
chemistry synthesis.
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
The initial synthesis of the aldehyde intermediate 519 was accomplished with a five-step reaction
sequence (Scheme 2) starting from 3,3-dimethylcyclohexanone 6. First, the ketone 6 was
converted to 7 by a Claisen condensation with dimethyl carbonate. Intermediate 7 was then
converted to vinyl triflate intermediate 8 which underwent a Suzuki cross-coupling reaction with
boronic acid 9 to produce the ester intermediate 10. Reduction of the ester 10 with LiBH4 led to
the alcohol 11 which was oxidized to aldehyde 5 using Dess-Martin periodinane. Silica gel
column chromatography was required to purify the aldehyde 5 to achieve the desired purity
profile.
Scheme 2. Medicinal Chemistry Route for Aldehyde 5
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 49
1
2
3
O
4
5
6
7
NaH, (MeO)2CO
MeO
NaH, Tf2O
O
THF, reflux
95%
DCM, -78 °C to RT
95%
O
6
7
8
9
O
O
B(OH)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
MeO
MeO
Cl
9
TfO
Pd(PPh3)4, CsF
DME/MeOH
8
Cl
10
70%
OH
O
LiBH4
Dess-Martin
Ether/MeOH
RT
DCM
85%
80%
Cl
Cl
11
5
A two-step, one-pot synthesis of aldehyde intermediate 5 (Scheme 3)20 was developed as a
streamlined alternative. Starting with the same starting material 3,3-dimethylcyclohexanone 6,
the ketone was converted to the chloro-aldehyde intermediate 12, which was directly subjected to
the Suzuki coupling with boronic acid 9 without need for isolation. Aldehyde 5 was isolated in
37% yield over the two steps after column chromatography purification.
Scheme 3. A New Synthesis for Aldehyde 5
OH
B
OH
O
15
1.0 eq.
O
1.6 eq. DMF
1.5 eq. POCl3
Cl
O
Pd(OAc)2/Bu4NBr
CH3CN/aq. K2CO3
37% (2 steps)
column chromatography
CH2Cl2
Cl
12
26
18
Cl
Although the synthesis of the aldehyde 5 was significantly improved from the five-step
sequence, its isolation and purification remained challenging due to its syrup-like physical state,
which made removal of process impurities difficult. The Suzuki cross-coupling 21 reactions for
both the initial and improved syntheses introduced several highly toxic mutagenic and
carcinogenic impurities (9, 13 and 14 in Figure 2) which required control to part per million
6
ACS Paragon Plus Environment

Page 7 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
(ppm) and part per billion (ppb) levels in drug substance, respectively.22 Consequently,
purification to reduce these impurities to desirable levels required tedious column
chromatography.
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
Cl
Cl
Cl
Cl
B(OH)2
9
13
14
Mutagenic
Carcinogenic
Figure 2. Mutagenic and Carcinogenic Impurities Generated from the Aldehyde Synthesis
In an effort to eliminate the chromatographic purification, attempts were made to form a
crystalline bisulfite adduct of aldehyde 5. Although crystallization of the bisulfite adduct was
possible, multiple crystallizations were required to achieve the desired purity profile, resulting in
substantial product losses. The significantly higher isolated yield of 5 from chromatography led
us to proceed with this method of purification in spite of the resource and time-intensity required.
The initial SNAr reaction involved reacting methyl 2,4-difluorobenzoate 15 with N-
triisopropylsilyl-protected hydroxylazaindole 16 (Scheme 4).23 Investigation of this reaction
revealed that substantial levels of desilylation occurred during the reaction and produced a
mixture of products with and without the TIPS protecting group. The mixture of products could
be converted to the desired intermediate 4a by treatment with TBAF. The major byproducts were
17 and 18, generated from non-regioselective displacement of the 4-fluoro group; interestingly,
only a minimal amounts of byproduct 19, from N-arylation of the deprotected azaindole, was
observed.
Scheme 4. Initial SNAr Reaction Using the TIPS-Protected Azaindole 16
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 49
1
2
3
4
5
6
7
8
9
CO2Me
F
O
OMe
F
O
OMe
O
1) K3PO4
diglyme
110 °C, 24 h
O
N
N
H
F
F
15
N
N
4a
H
2). TBAF/THF
60%
17
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
+
HO
CO2Me
O
CO2Me
O
N
N
TIPS
16
N
H
N
N
N
F
O
CO2Me
F
N
H
N
19
18
Based on this observation, we reasoned that the protection/deprotection steps could be
completely avoided and the SNAr reaction could be performed using the unprotected azaindole
20 (Scheme 5). Conditions for the SNAr with azaindole 20 were briefly screened; after evaluating
solvent, base and reaction temperature, it was found that the reaction conditions using a
combined solvent system of 2-Me-THF/DMF (10:1) with KOt-Bu afforded comparable results to
the original chemistry using TIPS-protected 16.
Scheme 5. The New SNAr Using unprotected azaindole 20
CO2Me
F
CO2Me
O
HO
10:1 2-Me-THF/DMF
KOt-Bu, 80 °C
+
N
N
N
H
N
30-60%
H
20
4a
F
F
1.05 eq.
15
During the scale-up of this SNAr reaction, the product purity profile was discovered to be
extremely sensitive to agitation and is heavily scale-dependent, because the reaction mixture
started off as a heterogeneous mixture due to the low solubility of 20. On 2-kg scale, the reaction
proceeded similarly to lab experiments in respect to the yield and product purity profile.
However, when the reaction scale was increased 10-fold, a substantial increase of the undesired
8
ACS Paragon Plus Environment

Page 9 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
byproducts was observed, which are from the SNAr at the 4- position and the double SNAr at
both 2 and 4-positionresulting in a significantly diminished yield of 33%. Chemical engineering
studies to understand the agitation power in relation to the reactor geometry was modeled, but
reproducibility challenges with respect to product purity profile and yield persisted.
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
The second SNAr reaction of 4a with piperazine 21 (Scheme 6) for the preparation of
intermediate 22a was initially carried out at 100 °C in DMSO,24 but the safety evaluation of this
reaction revealed that exothermic decomposition of the reaction mixture occurs at 200 °C with an
onset temperature at 130 °C. To address this safety concern for manufacture, the reaction
temperature could be reduced to 50 °C while still achieving full reaction conversion in ≤10 h.
Scheme 6. The SNAr Reaction for Intermediate 22a
H
N
CO2Me
CO2Me
O
N
O
21
H
8.0 eq.
N
N
H
N
DMSO
50 °C
N
N
H
22a
4a
F
89%
N
H
As expected, the dimeric impurity 22b (Figure 3) was also generated during the reaction and
needed to be controlled to <0.5% in the isolation of 22a due to limited purge in the downstream
steps. The initial conditions using 4 equivalents of piperazine 21 produced about 3% of dimer
with about 3-4% of starting material 4a remaining. When the amount of piperazine was
increased to 8 equivalents, the amount of the dimer was significantly reduced to ~0.4%; further
increasing the equivalents of piperazine diminished the amount of dimer formed, but was not
necessary as the target limit of the impurity was achieved.
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 49
1
2
3
4
MeO2C
N
N
CO2Me
5
6
O
O
7
8
N
N
9
NH
HN
22b
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 3. The Dimer Impurity from the Step 1 of Scheme 4
Minimal process optimization effort was directed to the reductive amination for the preparation
of 23a as the initial conditions were efficient. Coupling of piperazine derivative 22a and
aldehyde 5 was achieved using sodium triacetoxyborohydride at 20 °C to give 82% yield of 23a.
Notably, the reaction required more than10 h to achieve full conversion when <1.3 equivalents of
NaBH(OAc)3 was used; however, while increasing the amount of NaBH(OAc)3 did reduce the
reaction time, it had the undesired effect of increasing the levels of competitive aldehyde
reduction to alcohol 11.
The sulfonamide 3 was initially prepared in Discovery Chemistry by the SNAr reaction of 4-
fluoro-3-nitrobenzenesulfonamide 24a with the THP-amine 25 (Step 4 of Scheme 7).25 The
reaction proceeded well and was carried out at ambient temperature in MeCN. Although bulk
supplies of 24a were unavailable, it could be readily replaced with the less expensive chloro
analogue 24b, but heating to 80 °C was required to effect clean SNAr (Step 4 of Scheme 4).20
Since both 24a and 24b are strong sensitizing and mutagenic agents, an excess 1.4 equivalents of
the amine 25 was used to ensure complete consumption of the starting material.
The saponification of the methyl ester 23a to the acid 2 and the final coupling reaction of 2 with
the sulfonamide 3 with the core acid 2 (Step 5 of Scheme 7)26 were briefly studied and the initial
conditions were used without major modification.
10
ACS Paragon Plus Environment

Page 11 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
These initial process optimization efforts led to the identification of alternative chemistry for the
synthesis of key intermediates with improved scalability and efficiency, but more importantly
enabled rapid delivery of API. Using this route (Scheme 7), multi-kilogram quantities of drug
substance 1 were prepared, allowing early development of this promising drug candidate to move
forward without delay.
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
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 49
1
2
3
4
Scheme 7. Summary of the First Generation Large-Scale Synthesis of 1
5
6
7
8
CO2Me
O
CO2Me
HO
10:1 2-Me-THF/DMF
1.05 eq.KOt-Bu, 80 °C
F
N
N
H
+
9
4a
F
N
N
H
H
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
30-60%
Step A
20
15
F
8.0 eq.
21
DMSO
50 °C
1.05eq.
N
H
Step 1
89%
CO2Me
O
OH
B
OH
O
1.6 eq DMF,
1.5 eq. POCl3
CH2Cl2
O
N
Cl
1.0 eq.
N
H
9
N
O
22a
Pd(OAc)2/Bu4NBr
Cl
Step B
6
12
CH3CN/aq. K2CO3
37% (2 steps)
column chromatography
N
H
5
Cl
1.07 eq.
Step C
NaBH(OAc)3
THF, RT
Step 2
82%
CO2Me
CO2H
O
O
NO2
H
Step 3
N
LiOH
2.5 eq.
N
H
N
N
O
N
H
N
N
O
S
N
N
THF:EtOH:H2O
50 °C
O
NH
O
Step 5
90%
O
1). EDAC, DMAP
CH2Cl2
N
N
H
N
N
Et3N
23a
2
Cl
Cl
2. DMEDA
84%
NO2
NO2
O
H
X
N
H2N
O
O
1.44 eq.
S
NH2
O
25
O
S
Cl
1
NH2
1.0 eq.
O
CH3CN, DIPEA
80 °C
24a: X = F
24b: X = Cl
3
88%
Step 4
3. RESULTS AND DISCUSSION
Redesign of synthetic route: Following the development of the First Generation Synthesis, the
project timeline was significantly accelerated, allowing Process Chemistry only ten months to
develop a more scalable and robust synthesis to produce the primary stability batches. 27
12
ACS Paragon Plus Environment

Page 13 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
To develop a new synthesis under this compressed timeline, several key considerations for an
optimized manufacturing process had to be assessed and prioritized. Reaction safety, process
robustness, cost and labor effectiveness, facile impurity rejection, and environmental impact
were the most important considerations. Another key aspect of developing a scalable synthesis is
the strategic designation of regulatory starting materials (RSM);33 the RSMs should have good
chemical stability and desirable physical properties with a well-defined impurity profile and
should not be a major contributor to impurities in the API.
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
The first generation synthesis (Scheme 7) did not meet these criteria due to several major
concerns: (1) low overall yield and throughput resulting in high cost of drug substance, (2) poor
robustness of the SNAr reaction of hydroxyazaindole to produce 4a due to its poor
regioselectivity and extreme scale sensitivity, and (3) poor physical properties of the aldehyde
intermediate 5 complicating its isolation and the removal of mutagenic and carcinogenic
impurities from the Suzuki coupling.
In considering a more efficient and robust route, we wanted to preserve the sulfonamide coupling
as the final step since some process chemistry efforts had already been dedicated to optimizing
the process to ensure isolation of high purity API 1. We thought it would be advantageous to
leverage the new 2-step synthesis developed for aldehyde intermediate 5 ( Scheme 3). It was also
imperative for us to avoid isolation of aldehyde 5 in order to overcome the labor-intensive
cumbersome purification and isolation. Finally, we needed to address the robustness concerns of
the SNAr reaction of hydroxyazaindole 20 (Scheme 5) for its poor predictability and
reproducibility in producing intermediate 4a on scale.
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 49
1
2
3
4
5
6
7
8
9
With these objectives in mind, retrosynthetic analysis of the core ester 23a was performed
(Scheme 8). The key disconnection involved assembly of ester 23a either by a Buchwald-
Hartwig amination between the intermediate 4b and 26 or a SNAr reaction of 4a with 26. The
new piperazine intermediate 26 could be prepared from the reductive amination of aldehyde 5
with N-Boc-piperazine followed by boc-deprotection. This new reaction sequence could avoid
the isolation of aldehyde intermediate 5 and potentially furnish a crystalline intermediate.
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 8. Retrosynthetic Analysis of the Proposed New Synthesis
CO2Me
C-O Coupling
OH
Br
CO2Me
O
+
N
H
N
F
31
30
CO2Me
N
O
N
H
X
4a:X = F
4b: X = Br
N
N
CO2Me
F
H
SNAr Reaction
HO
N
N
+
+
N
H
N
H
N
Br
20
32a
O
N
Cl
23a
O
Cl
6
26
Cl
5
The 4-fluorobiaryl ether 4a could arise from a C-O cross-coupling reaction of the 5-
bromoazaindole 30 with 2-hydroxy-4-fluorobenzoate 31. Alternatively, the 4-bromobiarylether
4b could be prepared by a regioselective SNAr reaction of methyl 2-fluoro-4-bromobenzoate 32a
with 5-hydroxyazaindole 20. In this case, the cost and availability of 32a would have to be
addressed as it was expensive and unavailable in bulk quantities.
Proof-of-concept studies for the new route started with the evaluation of the alternate synthesis
for 4a. A variety of C-O cross-coupling reaction of 30 and 31 with both palladium and copper
14
ACS Paragon Plus Environment

Page 15 of 49
The Journal of Organic Chemistry
1
2
3
4
catalysts were evaluated and resulted in no desired product formation (Scheme 9). 28 We
5
6
attributed the poor reactivity to the poor nucleophilicity of electron-deficient phenol 31.
7
8
9
Scheme 9. Alternate Synthesis for Biarylether 4a via C-O Cross-Coupling
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
OMe
OH
O
OMe
O
Pd and Cu
catalysts
Br
X
+
N
N
N
N
H
H
F
F
4a
31
30
We then turned our attention to evaluating the alternate regioselective SNAr reaction of 32a and
20 (Scheme 10). With the introduction of the bromide at the 4- position, differentiating its
reactivity from the 2-fluoro, we no longer encountered any regioselectivity issues. In fact, the
SNAr reaction of 5-hydroxyazaindole 20 with 32a using NaOt-Bu in DMF20 proceeded cleanly
and the desired product 4b was obtained in high yield (>86%) as a crystalline solid.
Scheme 10. New SNAr Reaction for Bromobiarylether 4b
OMe
O
OMe
F
O
O
HO
NaOt-Bu / DMF
+
50 °C
>86%
N
N
N
N
H
H
Br
32a
Br
20
4b
A new reaction sequence for the preparation of 26 that avoids the isolation of aldehyde 5 was
evaluated (Scheme 11). The crude aldehyde 5 underwent clean reductive amination with N-Boc-
piperazine 28 to form the piperazine intermediate 29 in 74% yield over 3 steps with high purity
(>98%).20 The desirable physical property of 29 as a crystalline solid enabled thousand-fold
reduction of the mutagenic and carcinogenic impurities (9, 13 and 14 in Figure 2) by
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 49
1
2
3
4
5
6
7
8
9
crystallization. Additionally, deprotection of the Boc group using an aqueous hydrochloric acid
solution produced chloropiperazine 26 as a crystalline bis-HCl salt in 96% yield, further enabling
the purge of process impurities. The new reaction sequence for 26 produces higher overall yield
and affords robust control for rejection of impurities, but most notably avoided the need to
handle and isolate the syrup-like aldehyde intermediate 5 which had previously required tedious
column chromatography purification.
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 11. Alternate Synthesis of Chloropiperazine 26
O
B(OH)2
1.6 eq.DMF
1.5 eq.POCl3
CH2Cl2
O
1.0 eq
Cl
9
O
45 °C, 16h
Cl
Pd(OAc)2/Bu4NBr
CH3CN/aq. K2CO3
6
12
Cl
5
Boc
N
Boc
N
H
N
2HCl
N
H
1.2 eq
N
N
HCl, 60 °C
28
IPA
96% yield
Cl
NaBH(OAc)3
THF, RT
74% yield (3 steps)
Cl
29
26
The new syntheses developed for the key building blocks 4b and 26 provided sufficient
quantities of materials for investigation of the Buchwald-Hartwig amination.29 The coupling was
initially evaluated with a variety of catalysts with methyl ester 4b; however, C-N bond formation
was relatively inefficient and substantial amounts of byproducts were formed. The two major
byproducts were the piperazine amide from transamidation and the 4-bromobenzoic acid from
saponification of the methyl ester. A minor amination byproduct was the result of C-N coupling
at the aryl chloride, indicating there could be potential chemoselectivity issues (i.e. oxidative
16
ACS Paragon Plus Environment

Page 17 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
addition into Br vs Cl). Observation of the ester-related byproducts from the initial set of
experiments led us to switch to the more sterically hindered tert-butyl ester 4c, which we
hypothesized should be less prone to transamidation as well as saponification. Indeed, when tert-
butyl ester 4c was used in the coupling reaction, very little transamidation and saponification
were observed. Our most promising initial result was with a high catalyst load of [Pd(P(t-
Bu)3)Br]2 (20 mol% of total Pd),29-30 the reaction proceeded with high conversion (>99%) and
high yield (>93%). It was found that >3 equivalents of NaOt-Bu were required (Scheme 12).
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 12. Initial Buchwald-Hartwig Amination for 23c
CO2t-Bu
O
H
N
N
H
N
CO2t-Bu
N
N
10 mol% [(Pt-Bu3)PdBr]2
O
N
+
N
N
H
3.2 equiv NaOt-Bu
THF, 30 °C
Br
4c
Cl
23c
26-Free base
Cl
To develop a cost effective manufacturing process, the attempts were made to lower the Pd
catalyst loading for the C-N coupling. When the amount of [Pd(P(t-Bu)3)Br]2 used was reduced
to 1 mol% (for 2 mol% of total Pd), the reaction conversion stalled out at ~40% when 3.2
equivalentsof NaOt-Bu were used (Figure 4). However, full conversion (>99%) could be
achieved with larger excesses of base (~7 equivalents); we noted that less NaOt-Bu was required
with higher Pd loadings. At present, we hypothesize this phenomenon is a result of unidentified,
but subtle organometallic interactions;.the -basic N-1 and N-7 of the azaindole may complex to
Pd to form catalytically inactive, off-cycle Pd—N species and the excess NaOt-Bu may be
necessary to recover the catalyst via a Pd alkoxide.31 Previously reported methods for
17
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 49
1
2
3
4
5
6
7
8
9
aminations with LiHMDS in the presence of similar unprotected heterocycles gave poor
conversion in our case.30e-f Use of a large excess of NaOt-Bu in this coupling reaction has little
impact to the overall cost of goods as NaOt-Bu is relatively inexpensive and could be readily
sourced.
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
100.0
75.0
50.0
25.0
2% Pd
4% Pd
0.0
0
1
2
3
4
5
6
7
8
9
NaOt-Bu (equiv)
Figure 4. Impact of NaOt-Bu Equivalents on C-N Coupling Conversion
The final saponification of the tert-butyl ester 23c turned out to be more challenging than
expected, the previous saponification conditions (LiOH in 1:1:1 ratio of THF/EtOH/H2O at 50
°C) were ineffective for 23c. The reaction was extremely slow with very little conversion even
with extended reaction times. When the reaction temperature was raised, significant amounts of
degradation were observed. Other commonly used basic saponification conditions were also
examined without success. Switching to acid-mediated conditions with trifluoroacetic acid in
dichloromethane or phosphoric acid in acetonitrile gave low yields of product and was
accompanied by substantial amounts of decarboxylation. Alternatively, it was found that the tert-
butyl ester 23c could be cleanly converted to the methyl ester 23a under acidic conditions, which
subsequently undergoes clean saponification to the desired acid 2 as described before (Step 3,
18
ACS Paragon Plus Environment

Page 19 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
Scheme 7). Although both reactions proceeded cleanly, this was non-ideal due to the additional
step required.
7
8
9
Our analysis of the literature revealed that “anhydrous hydroxide,” generated from the reaction
of KOt-Bu with a carefully controlled stoichiometry of water, could be used to carry out the
saponification of a variety of bulky esters in organic solvent.32 We thought this under-utilized
procedure could be promising; and in fact, we observed clean, but incomplete saponification
when 23c was treated under the reported conditions using 2 equivalents of KOt-Bu and 1
equivalents of water in THF at room temperature. After some optimization, the reaction
proceeded to completion to give acid 2 in >90% yield when the saponification was performed
with 5 equivalents of KOt-Bu and 3 equivalents of water at 55 °C (Scheme 13).
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 13. Saponification of t-Butyl Ester with “Anhydrous hydroxide”
OH
O
O
Ot-Bu
O
O
N
N
N
N
H
H
N
N
KOt-Bu, H2O
N
N
THF
55°C
>90%
Cl
Cl
23c
2
Having demonstrated that the Buchwald-Hartwig coupling could be used to successfully prepare
23c, a cost effective synthesis of key intermediate 32c was required. As previously mentioned,
methyl benzoate 32a was expensive and unavailable in large quantities and this was also true for
the tert-butyl analogue 32c. So, a new synthesis for 32c that utilizes inexpensive and readily
available raw material 33 was developed (Scheme 14).
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 49
1
2
3
4
5
6
7
8
9
Selective metalation of 4-bromo-2-fluoro-1-iodobenzene 33 with iPrMgCl affords Grignard
intermediate, which was then reacted with Boc2O at -5 °C to produce the desired tert-butyl ester
32c in 88% yield. A simple aqueous workup afforded product in >98% area purity, which could
be used in the following SNAr reaction without further purification. The timely development of
this straightforward and economical synthesis of 32c provided a reliable source of the starting
material and was an important factor in deciding to pursue this new synthetic 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
Scheme 14. New Synthesis for the tert-Butyl Ester 32c
I
MgCl
O
Ot-Bu
F
F
1.3 eq. iPrMgCl
5 °C
F
1.5 eq. Boc2O
88 %
THF
Br
Br
Br
32c
33
Grignard
Intermediate
These proof-of-concept studies demonstrated the improved efficiency and convergence of the
new synthetic route. A Buchwald-Hartwig amination reaction was used to connect the two key
building blocks bromobenzoate 4c and chloropiperazine 26. Moreover, highly scalable
processes for both 4c and 26 were implemented. This new synthesis addressed all of the
aforementioned challenges associated with the first-generation process, while also importantly
reducing the cost of goods. Having achieved the goals set forth, the decision was made to further
optimize this route to synthesize 1.
Process Development of the New Synthesis. Robust control strategies that define the process
parameters and operating ranges must be established to ensure consistent process performance
and product quality. They are based upon thorough examination and fundamental understanding
of each reaction and the corresponding risk assessment. Control of material attributes and purity
20
ACS Paragon Plus Environment

Page 21 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
profiles for each intermediate and the RSMs33 are also important parts of the overall strategy and
are established based on process understanding of impurity formation and their downstream fate.
7
8
9
The development of a robust manufacture process for 1 was focused on further optimization and
understanding of the reactions starting from the designated RSMs (4c, 26 and 24b). One of the
more important aspects of this endeavor was to optimize and study the Buchwald-Hartwig
amination between 4c and 26. The first step was to identify a more efficient catalyst. Based on
the promising initial result with the [Pd(P(t-Bu)3)Br]2 precatalyst, a focused ligand screen was
performed by evaluating tert-butyl-substituted and other bulky phosphines (Figure 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
PtBu2
N
PR2
N
O
O
PtBu2
PtBu3
Ph
Ph
O
P
Ph
N
N
NMe2
Ph
C
D
: R=tBu
: R= Cy
F
A
B
E
(Amphos)
PtBu2
PtBu2
P(Ad)2
tBuPCy2
MePtBu2
BnP(Ad)2
Fe
N
O
G
H
I
J
K
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 49
1
2
3
4
Figure 5. Focused Ligand Screen for C-N coupling
5
6
7
8
9
From this ligand screen where 2.0 mol% of total Pd was used, not surprisingly, PtBu3 (A in
Figure 5) was still effective, but both Amphos (F) and DiTBPF (I) were also observed to be
effective for the coupling of 4c and 26, each giving >80% conversion to 23c. In order to better
differentiate among these three ligands, the amination was performed under more challenging
conditions, now using only 1.0 mol% of total Pd (Figure 6). Under these stressed conditions,
Amphos29g was discovered to be the superior ligand (entries F and G in Figure 6). Although
[Pd(π-cinnamyl)Cl]2 gave a slightly more active catalyst than Pd2dba3, due to cost and
commercial availability, the decision was made to move forward with the process using Pd2dba3.
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
B
C
D
E
F
[(PtBu3)PdBr]2
[-cinnamyl)PdCl]2 + PtBu3
Pd(DiTBPF)Cl2
Pd2dba3 + DiTBPF
Pd(OAc)2 + PhB(OH)2 + DiTBPF
[-cinnamyl)PdCl]2 + AmPhos
Pd2dba3 + AmPhos
G
H
Pd(OAc)2 + PhB(OH)2 + AmPhos
Figure 6. Evaluation of Pd Precursors for the Amination
The starting material charge ratio of benzoate 4c to chloropiperazine 26 was found to be
important with respect to formation of impurities. When the chloropiperazine 26 was used in
excess, increased levels of Cl-amination impurity 37 was generated from further coupling of 26
with the desired product 23c (Scheme 15). Although the Pd catalyst reacts preferentially with the
aryl bromide at low conversion, while approaching high conversion of bromobenzoate 4c, the
22
ACS Paragon Plus Environment

Page 23 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
low concentration of aryl bromide relative to the chloride in 23c leads to the undesired coupling.
Therefore 4c was used in slight excess (1.04 equivalents) to ensure that the impurity 37 is formed
in acceptable level as impurity 37 and its daughter impurities are not readily purged downstream.
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 15. The Key Impurity Generated from the Amination
CO2t-Bu
O
CO2t-Bu
O
H
N
N
1 mol Pd2(dba)3
2.2 mol Amphos
5 eq. NaOt-Bu
THF, 55 °C
N
H
N
N
N
N
H
N
N
N
+
Cl
N
26-Free base
N
23c
Cl
37
Cl
As previously observed, efficient cross-coupling was dependent on the addition of excess
equivalents of NaOt-Bu and this was also found to be the case with the Pd/Amphos catalyst.
However, with 1.5 mol% of the Pd/Amphos catalyst as few as 4 equivalents were necessary to
achieve full conversion. It was also observed that the quality of NaOt-Bu had a significant
impact on the reaction performance as such large amount is used. When old lots of NaOt-Bu
were used, the reaction proceeded to generate high levels of benzoic acid impurities, from the
saponification of both 4c and 23c. We assumed that the old lots of base contained higher level of
NaOH generated from quenching of NaOt-Bu with atmospheric moisture, analogous to the
“anhydrous hydroxide” generated in the saponification of the tert-butyl ester 23c to the core acid
2. Saponification of 4c generates a sodium carboxylate which in theory could couple with
chloropiperazine 26 to give core acid 2 directly; however, due to its poor solubility in THF, the
carboxylate does not undergo this amination, so the saponification of 4c is an unproductive
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 49
1
2
3
4
5
6
7
8
9
pathway resulting in lower yield of 23c and higher levels of impurities. Analytical control of the
quality of NaOt-Bu to ensure low levels of NaOH would be challenging and impractical,
especially given that the base would be re-exposed to atmospheric conditions following the
quality testing. Therefore, it was necessary to develop a robust operational control strategy to
ensure minimal amounts of NaOH would be introduced to the reaction. Given the very low
solubility of NaOH in THF, filtration of the NaOt-Bu/THF solutions to remove the sodium
hydroxide in laboratory experiments demonstrated that the amination reaction proceeded
consistently with lots of varying quality. Incorporation of a filtration unit operation during
manufacturing for removal of NaOH was straightforward and offered a practical and robust
strategy to enable this sensitive reaction to be carried out reproducibly.
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
The sensitivity of the coupling reaction to the level of oxygen in the system could adversely
affect the reaction performance, producing variable results. This potential concern was addressed
by controlling and monitoring the oxygen level of the headspace using an oxygen monitor. It was
determined that an oxygen level of ≤40 ppm in the headspace was sufficient to ensure consistent
reaction performance and could be achieved by sub-surface sparging of solvents with dry
nitrogen and active exchange of reactor atmospheres with vacuum/N2-purge cycles. With control
of the oxygen level, under the optimized conditions (6.5 equivaletns NaOt-Bu, 0.8 mol%
Pd2dba3, 1.6 mol% Amphos in THF) the amination proceeded in >99% conversion with about
90% isolated yield following aqueous workup and crystallization.
The saponification using the “anhydrous hydroxide” was also further studied and optimized.
Ultimately, we chose 2-Me-THF as the reaction solvent because it is less hygroscopic than THF,
providing better control of water levels in the reaction. In addition, the aqueous work-up and
crystallization purification could be performed more effectively in 2-Me-THF. When a series of
24
ACS Paragon Plus Environment

Page 25 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
saponification experiments were performed with 5 equivalents of KOt-Bu and varying amounts
of water (Figure 7), the reaction proceeded to completion only when the ratio of water to KOt-Bu
was ≤1 (i.e. an excess of KOt-Bu relative to water). When water was used in excess relative to
KOt-Bu, incomplete reactions were observed. The optimal molar ratio was determined to be 0.6
using 5 equivalents of KOt-Bu.
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
100
80
60
40
20
0
0
1
2
3
4
5
6
7
Water equivalent
Figure 7. Impact of Water Amount Using 5 equivalents of KOtBu to Reaction Conversion
Higher temperatures drastically reduced the reaction time; the saponification proceeded to
completion cleanly within 2 h at 50 °C, while only 43% conversion was achieved during that
time period at 21 °C. Since there was no detrimental impact on the impurity profile, the reaction
was performed at 50 °C to keep the reaction cycle time low. Under the optimized conditions (6
equivalents KOt-Bu and 3.6 equiv H2O at 50 °C), the reaction proceeded with >98% conversion
and ultimately afforded about 90% isolated yield.
Developing robust strategies for the final sulfonamide coupling reaction and isolation was
critical to ensure the API quality. The low solubility of the API (1) in most organic solvents
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 49
1
2
3
4
5
6
7
8
9
limited the solvent options for the amide coupling and isolation. Moreover, the poor
nucleophilicity of the sulfonamide resulted in a slow reaction and competing byproduct
formation. Screening of coupling of reagents (e.g. CDI, HATU, T3P and EDAC) and different
amine bases (e.g. Hünig’s base, Et3N, DMAP, DBU, DABCO, pyridine, collidine, lutidine, and
N-methylmorpholine) confirmed that the initial conditions using EDAC and DMAP/Et3N
produced the best results with respect to conversion and purity. Solvent screening revealed that
dichloromethane was the optimal solvent and it was also found that the coupling was best run at
room temperature. Although increased temperatures (35 °C to reflux) decreased the reaction
times, it also increased the levels of impurities.
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
Efforts were also focused on understanding the formation of impurities and their rejection.
Several process impurities were identified and their structures were determined using 2D NMR
and MS analyses (Figure 8); the low level impurities 38, 39 and 40 could be readily removed
during aqueous work-up and crystallization. The major bis-amide 41 impurity was difficult to
reject and its level varied between batches; it slowly increased during the reaction, meanwhile
the reaction rate for forming 1 decreased.
The order of addition of the reagents also impacted the bis-amide impurity level. When EDAC
was added last to a suspension of the sulfonamide 3, core acid 2 and DMAP/Et3N, the bis-amide
41 was formed in 8%. In contrast, slowly adding core acid 2 to the reaction mixture last resulted
in only 5% of the bis-amide. The addition time of the core acid 2 solution also played a key role;
longer addition times (8 h) led to incomplete reactions with up to 9.0% of core acid 2 remaining.
After a series of experiments (Figure 9), it was determined that the best reaction profile was
obtained with slow addition of a mixture of core acid 2 with Et3N in dichloromethane to a
26
ACS Paragon Plus Environment

Page 27 of 49
The Journal of Organic Chemistry
1
2
3
4
suspension of sulfonamide 3, DMAP and EDAC in dichloromethane with an optimal addition
5
6
time of 2~6 h.
7
8
9
O
O-
N+
H
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
O
S
NH
NO2
O
N
H
N
H
N
S
N
NO2
O
N
H
O
N
O
N
O
O
NH
HN
N
O
S
O
O
O
N
O
N
O
O
O
O
O
N
N
N
N
N
O
H
H
N
N
N
N
N
N
H
N
N
N
N
N
H
N
N
Cl
Cl
Cl
Cl
Cl
38
39
40
Bis-amide
41
Figure 8. Impurities Generated in the Sulfonamide Coupling Reaction
Figure 9. Impact of the Addition Time of the Core Acid (2) Solution
It was also discovered that treatment of bis-amide 41 with N,N-dimethylethylediamine
(DMEDA) led to the re-formation of desired product 1 along with the corresponding DMEDA-
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 49
1
2
3
4
amide impurity 42 which was readily rejected (Scheme 16). This discovery further ensured the
5
6
process robustness with respect to impurity rejection.
7
8
9
Scheme 16. Derivatization of the Major Dimer API Impurity
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
O
O-
N+
H
N
O
NO2
N
H
O
S
N
O
NH
O
O
O
S
O
NH
O
NH
O
O
H2N
2.5 eq.
O
N
N
N
O
N
N
N
N
H
N
N
O
H
N
N
N
N
+
N
N
H
N
N
Cl
Cl
Cl
1
Bis-amide 41
DMEDA-amide 42
Cl
The API 1 was found to readily form solvates which had low solubility, making the work-up of
the amide coupling product challenging. After examination of different solvents and solvent
mixtures, we opted to use dichloromethane for the work-up and isolation. A dichloromethane
solvate was found to spontaneously crystallize from solution during aqueous extractions;
therefore the work-up was performed at high dilution (20 L/kg) and at 35 °C to keep the API
soluble. In addition, a small amount of methanol (0.5 L/kg) was added to the organic layer after
each extraction to increase product solubility and prevent product crystallization. The extractive
work-up uses aqueous acetic acid to remove excess DMAP followed by sodium bicarbonate
wash to remove residual acetic acid and aqueous wash with 5% NaCl to prevent emulsions.
The final product purification by crystallization effectively rejected product-related impurities
and consistently produced 1 in acceptable yield and high quality with desired solid polymorph
property. It was achieved by slowly adding ethyl acetate as anti-solvent to attain a 10/1/9 mixture
28
ACS Paragon Plus Environment

Page 29 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
of dichloromethane/methanol/ethyl acetate, with a targeted total volume of 16 L/kg. As
previously mentioned, the addition of methanol (1:10 ratio to dichloromethane) increases the
product solubility. Additionally, the presence of methanol also increases the solubility of the
reaction impurities and resulted in a robust purification with rejection of up to 5% DMEDA-
amide. This optimized process (Scheme 17) has been used for producing 1 with consistent high
quality.
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 17. Optimized New Synthesis of 120
OH
Boc
N
H
N
Boc
N
B
O
OH
2HCl
1.6 eq.DMF
1.5 eq.POCl3
9
O
N
Cl
N
H
N
HCl, 60 °C
28
1.0 eq.
CH2Cl2
1.2 eq.
O
IPA
Pd(OAc)2/Bu4NBr
CH3CN/aq. K2CO3
5
Cl
1.2 eq. NaBH(OAc)3
THF, 22 °C
Cl
Step B
6
96%
12
29
Cl
26-2HCl
Cl
74% (3 steps)
1. K3PO4/Tol
H
N
I
CO2tBu
F
CO2tBu
F
HO
1.2 eq. NaOt-Bu
N
O
1)1.3 eq. iPrMgCl
50 °C
+
N
N
N
N
2) 1.5 eq Boc2O
-5 °C
DMF
86%
H
H
1.2 eq.
Br
Br
Br
4c
20
Cl
88 %
33
32c
26-Free Base
1.6 mol% Pd2(dba)3
1.6 mol % Amphos
6.5 eq.NaOt-Bu
THF, 55 °C
89%
NO2
H
N
CO2tBu
O
CO2H
O
O
O
S
N
N
O
NH
N
N
O
H
H
N
N
N
O
6.0 eq. KOt-Bu
4.1 eq. H2O
N
N
N
H
2-MeTHF, 50 °C
88%
N
N
1). EDAC, DMAP
CH2Cl2
Et3N
Cl
Cl
23c
2
2. DMEDA
84%
NO2
H
O
NO2
Cl
1. ABT-199
Venetoclax (1)
N
H2N
Cl
1.4 eq.
O
O
25
S
S
O
CH3CN, DIPEA
O
O
NH2
NH2
80 °C
88%
24b
1.0 eq.
3
29
ACS Paragon Plus Environment