The First Kilogram Synthesis of Beclabuvir, an HCV NS5B Polymerase Inhibitor

Article abstract of DOI:10.1021/acs.oprd.8b00214

The process development and kilogram-scale synthesis of beclabuvir (BMS-791325, 1) is described. The convergent synthesis features the use of asymmetric catalysis to generate a chiral cyclopropane fragment and coupling with an indole fragment via an alkyl

Full text of DOI:10.1021/acs.oprd.8b00214

Subscriber access provided by Karolinska Institutet, University Library
Full Paper
The First Kilogram Synthesis of Beclabuvir,
an HCV NS5B Polymerase Inhibitor
Jeffrey Bien, Akin H. Davulcu, Albert J DelMonte, Kenneth J. Fraunhoffer, Zhinong Gao, Chao Hang, Yi
Hsiao, Wenhao Hu, Kishta Katipally, Adam Littke, Aghogho Pedro, Yuping Qiu, Maria Sandoval, Richard L.
Schild, Michelle Soltani, Anthony Tedesco, Dale Vanyo, Purushotham Vemishetti, and Robert E. Waltermire
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.8b00214 • Publication Date (Web): 16 Aug 2018
Downloaded from http://pubs.acs.org on August 16, 2018
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 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The First Kilogram Synthesis of Beclabuvir, an
HCV NS5B Polymerase 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
Jeffrey Bien, Akin Davulcu, Albert DelMonte*, Kenneth J. Fraunhoffer, Zhinong Gao, Chao
Hang, Yi Hsiao, Wenhao Hu, Kishta Katipally, Adam Littke, Aghogho Pedro, Yuping Qiu,
Maria Sandoval, Richard Schild, Michelle Soltani, Anthony Tedesco, Dale Vanyo,
Purushotham Vemishetti, Robert E. Waltermire
Chemical and Synthetic Development, Bristol-Myers Squibb Company, One Squibb Drive,
P.O. Box 191, New Brunswick, New Jersey, 08903, United States
Albert.DelMonte@bms.com
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 44
1
2
3
4
Table of Contents graphic
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ACS Paragon Plus Environment

Page 3 of 44
Organic Process Research & Development
1
2
3
4
5
Abstract:
6
7
8
9
The process development and the kilogram-scale synthesis of Beclabuvir BMS-791325
(1) is described. The convergent synthesis features the use of asymmetric catalysis to
generate a chiral cyclopropane fragment and coupling with an indole fragment via an
alkylation. Subsequent palladium catalyzed intramolecular direct arylation efficiently
builds the central 7-membered ring. The target was prepared in 12 linear steps with 5
isolations in an overall yield of 8 %.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords: asymmetric, cyclopropanation, direct arylation, Palladium, Rhodium, catalysis
Introduction
The hepatitis C virus (HCV) is transmitted through infected blood and currently
affects an estimated 2–
3 percent of the world’s population.1 Infection with hepatitis C is
frequently asymptomatic but left untreated often leads to cirrhosis of the liver, liver failure,
hepatocellular cancer and even death. In the not too distant past, most infections were treated
with pegylated interferonꢀα and ribavirin. Treatment typically lasted 24–48 weeks and was
physically demanding as patients often experience fluꢀlike side effects. Fortunately, for those
afflicted with HCV, there have been a number of modern treatments that offer reduced side
effects and shortened treatment duration.2,3 In addition, many patients are treated with
combinations of these drugs and experience even greater cure rates. The HCV NS5B RNA
polymerase is an attractive target for drug inhibition as it is required for HCV replication. The
potential for a very selective inhibitor exists as there is no analogous enzyme in human cells.
It was for this purpose that Beclabuvir (BMSꢀ791325) (1) was selected as a potential drug
candidate for the inhibition of HCV NS5B RNA polymerase.4 This paper details the
development of a convergent chemical route and production of multiꢀkilogram quantities of 1,
which were required to meet initial toxicological and clinical supply material demands.
Results and Discussions
In order to support the clinical development, a short and efficient route was desired
which addressed the challenges of the Discovery route (Scheme 1). Although this approach
adequately met initial material requirements, a more convergent approach that would be
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 44
1
2
3
4
5
6
7
8
9
amenable to rapid scale up was desired. One aspiration for a scalable synthesis was to
eliminate the late stage achiral cyclopropanation and preparatory HPLC chiral separation as
this would greatly impact yield and throughput. There was also concern regarding use of the
reagent methyl 2ꢀ(dimethoxyphosphoryl)acrylate (9) utilized in the synthesis of 8, as we felt it
would require extensive development to understand and control potential polymerization
challenges.
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
4
ACS Paragon Plus Environment

Page 5 of 44
Organic Process Research & Development
1
2
3
4
Scheme 1. Discovery Synthesis of 1.a
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
`
aReagents and conditions: (a) THF, Et3SiH, cyclohexanone (77%); (b) Br2, HOAc (76%); (c) NaOH, THF,
MeOH (99%); (d) CDI, THF; (e) N,Nꢀdimethylsulfamide, DBU, THF (74%); (f) Pd(PPh3)4, LiCl, Na2CO3,
EtOH, toluene (88%); The arylboronic acid aldehyde is synthesized in 4 steps;5 (g) Cs2CO3, DMF (71)%; (h)
trimethylsulfoxonium iodide, NaH, DMSO (83%); (i) ChiracelꢀOJ chromatography, 44% recovery (50%
theoretical); (j) NaOH, THF, MeOH, H2O, (86%); (k) TBTU, DIPEA, DMF; (l) Chromatography, MeOH,
CH2Cl2, (70%).
New Route Strategy
Although a number of potential synthetic approaches were considered, we
concentrated our efforts on identification of a convergent route. As shown in the
retrosynthetic analysis (Figure 1), a disconnection strategy leading back to an achiral indole
fragment (16) and a chiral cyclopropane fragment (17) was envisioned.
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Revised Retrosynthetic Analysis of 1.
When considering the variety of options within this strategy, we leveraged existing
knowledge. Work by our Discovery colleagues suggested a final coupling of intermediate 13
with the bridged piperazine 14 was preferable, as they found this approach afforded 1 in
higher purity than alternative end games. An added benefit is that the relatively expensive
bridged piperazine 14 is introduced in the final step. Published work by Kosikowski6
(Scheme 2) also suggested that if the indole and cyclopropyl fragments were already coupled
through the carbonꢀnitrogen bond (as shown in intermediate 15), a direct intramolecular
palladium catalyzed direct arylation was possible to form the 7ꢀmembered ring and produce
compound 13. This was an important precedent as the direct arylation strategy would
significantly streamline the route by avoiding unnecessary steps to install a bromide on the
indole and a boronic ester on the aryl group.7
The alcohol moiety of the cyclopropyl
fragment had to be activated to function as an alkylating agent, and preparation of cyclopropyl
tosylate 17 not only served in this role, it also imparted a high level of crystallinity, which was
viewed to be advantageous at the time. Another approach to improve the convergent nature of
the synthesis was to couple the indole and cyclopropyl fragments with the N,Nꢀ
dimethylsulfamide moiety already installed on the indole 16, obviating the need for ester
hydrolysis and sulfamide coupling in the endꢀgame of the synthesis.
Scheme 2. Precedence of Direct Indole Arylation to Form a 7-Membered Ring
6
ACS Paragon Plus Environment

Page 7 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
There are several practical advantages to this synthetic approach: (1) the convergent nature
leads to a more efficient synthesis; (2) an earlier introduction of the chiral cyclopropane
results in higher throughput even if a resolution is required; (3) the opportunity to modify the
synthesis of the individual fragments to iteratively evolve the synthesis into a commercially
viable route without having to modify the overall strategy.
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 the Indole Fragment 16
Scheme 3 reveals our synthesis of the indole fragment 16. We adopted a reductive
alkylation and hydrolysis approach to prepare indole 18. A telescoped multiꢀstep process was
utilized to install the N,Nꢀdimethylsulfamide to afford 16.
The literature method for the reductive alkylation8 only afforded a 36% yield of 3
when the reaction was conducted with cyclohexanone. The original conditions were
optimized for order of addition, acid, temperature, stoichiometry, solvents, and crystallization
conditions. The modified process involved the controlled addition of TFA to a cold (0 ºC)
toluene mixture of 2, Et3SiH (3 equivalents), and cyclohexanone (2 equivalents) while
maintaining the temperature ≤20 ºC. The reaction was slowly warmed to 20ꢀ25 ºC and
monitored for reaction conversion. Upon reaction completion, the addition of petroleum ether
produced a suspension. Filtration, washing and drying of the cake directly afforded the
crystalline material. This process was demonstrated on pilot scale and afforded 89 kg of 3 in
64.7ꢀ69.9% yield and >99.6 HPLC area percent purity over 2 batches. These procedures were
repeated in a subsequent campaign and afforded 374 kg of 3 in 74.5ꢀ75.3% yield9 and >99.3
HPLC area percent purity over 4 batches.10
Scheme 3. Synthesis of Indole 16.
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The hydrolysis conditions for the conversion of 3 to 18 were optimized for
concentration of base, temperature, solvents, and crystallization conditions. The optimized
conditions required the treatment of 3 in 2 N NaOH solution (6 L/kg)11 and isopropanol (IPA)
(6 L/kg) at 75 ºC. Increased concentrations of NaOH lead to transesterification with IPA.
Upon reaction completion, the mixture was cooled to 50 ºC and neutralized with 2.5 N HCl to
pH ≤1 resulting in a white suspension. Cooling to 25 ºC and filtration afforded the desired
product 16. This process was demonstrated on scale and afforded 39.9 kg in 95.2ꢀ96.5% yield
and 99.8ꢀ99.9 HPLC area percent purity over 2 batches. These results were also duplicated in
a subsequent campaign and afforded 342 kg in 96.2ꢀ97.2% yield and 99.6ꢀ99.7 HPLC area
percent purity over 4 batches.
The coupling of 18 with sulfonamide proved more challenging than initially
anticipated. Although the desired product 16 was obtained when indole 18 was subjected to
CDI mediated coupling conditions analogous to those used successfully for the 2ꢀbromo
analog 5, unexpectedly high levels of impurity 22 (Figure 2) were also produced by a
competitive nucleophilic attack by the indole nitrogen on the activated intermediate 19.12
Although impurity 22 showed some promise of being able to convert to the desired indole 16
under relatively harsh conditions, (prolonged heating in the presence of excess N,Nꢀ
8
ACS Paragon Plus Environment

Page 9 of 44
Organic Process Research & Development
1
2
3
4
dimethylsulfonamide), we felt a more attractive approach would be to circumvent the
5
6
formation of impurity 22 entirely via an in-situ protection of the indole nitrogen.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Indole Coupling Impurity 22.
During the development of this chemistry, we found the optimal sequence involved
generation of the acyl imidazole 19 with CDI, followed by Boc protection. By introducing
catalytic 4ꢀdimethylaminopyridine (DMAP) (0.2 equivalents) during formation of 19, it was
found there was no initiation period and that the offꢀgassing was addition controlled which is
an important control feature from a safety perspective. Boc protection of 19 effectively led to
the formation of intermediate 20 which upon treatment with N,Nꢀdimethylsulfamide and 1,8ꢀ
diazabicyclo[5.4.0]undecꢀ7ꢀene (DBU) afforded intermediate 21. DBU was found to perform
better than triethylamine (TEA), diisopropylethylamine (DIPEA), Cs2CO3, pyridine and 1,4ꢀ
diazabicyclo[2.2.2]octane (DABCO). Although acidic conditions were effective at removing
the Boc protecting group, the associated CO2 offꢀgassing was a concern as it was not easily
controlled. However, basic conditions at elevated temperatures (65 °C) were effective at
removing the Boc group from 21 and allowed for the addition controlled evolution of CO2
during the subsequent neutralization with HCl. A number of solvents were screened for this
reaction sequence with dichloromethane (DCM), acetonitrile (ACN), tetrahydrofuran (THF),
dimethylformamide (DMF), and dimethylacetamide (DMAc) all demonstrated to be viable.
Ultimately, DCM was chosen as it allowed for a facile distillative swap to isopropanol for the
basic hydrolysis of the Boc group. Utilizing this approach, 38 kg of indole 16 was produced
in 98.9ꢀ99.3 HPLC area percent purity and 65.6ꢀ76.6% yield in two batches. These results
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 44
1
2
3
4
5
6
7
8
9
were also duplicated in a subsequent campaign in which 254 kg of 16 were prepared in 99.4ꢀ
99.7 HPLC area percent purity and an improved 76.3ꢀ77.4% yield in 4 batches.
Overall, the synthesis of the indole fragment 16 was conducted in 6 chemical
transformations, but only 3 isolations in 53 % overall 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 the Chiral Cyclopropane
Multiple potential approaches to the cyclopropane fragment were considered. During
our investigations, the acid alcohol cyclopropane 23 (Scheme 4) emerged as a key
intermediate in the synthesis of the tosylate 17. Intermediate 23 was not only highly
crystalline, but also formed a highly crystalline intermediate with (R)ꢀαꢀmethylbenzylamine
((R)ꢀAMBA) to afford 23•(R)-AMBA. These crystalline properties provided flexibility to the
synthetic approach to either leverage chiral salt formation as a way to conduct a resolution of
racemic 23 or as a tool to upgrade 23 of moderate enantiomeric excess (ee). The 23•(R)-
AMBA could then be converted to tosylate 17 via an esterification and subsequent tosylation
reaction (vide infra).
Scheme 4. Synthesis of Cyclopropane Fragment 17.
We envisioned that cyclopropane 23 could be produced by functional group
transformation of the vinyl cyclopropane 24 as shown retrosynthetically in Figure 3. The
seminal work by Huw Davies13 was clear precedent for the synthesis of the trisubstituted
cyclopropane 24 in good enantioselectivity from the tetrakis[1ꢀ[[4ꢀalkyl(C11ꢀ
C13)pheynyl]sulfonyl]ꢀ(2S)ꢀpyrrolidinecarboxylate]dirhodium (Rh2(SꢀDOSP)4) catalyzed
reaction between vinyl diazoacetate 25 and styrene 26.
10
ACS Paragon Plus Environment

Page 11 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Retrosynthetic Analysis to Cyclopropane 23
The synthesis of vinyl diazoester 25 involved reacting 27 with 4ꢀ
acetamidobenzenesulfonyl azide (4ꢀABSA) and DBU in ACN (Scheme 5). The crude
reaction mixture was diluted with methyl tertꢀbutyl ether (MTBE) and an aqueous workup
effectively removed the DBU byꢀproducts and ACN. This also induced the crystallization of
the sulfonamide byꢀproduct, which was removed via filtration. The mixture was solvent
swapped to nꢀheptane and afforded a 6–9 wt% solution of diazoester 25 in nꢀheptane in 75–80
% yield. The removal of ACN was found to be a critical operation as residual levels were
detrimental to catalyst performance in the cyclopropanation. Controlling ACN levels to ≤ 0.5
% (relative to the nꢀheptane solvent) ensured high quality product in the subsequent step.
Scheme 5. Synthesis of Vinyl Diazoacetate 25.
Before scaling up it was important to understand any stability or safety issues
associated with this specific vinyl diazoester 25. Fortuitously, different from most diazo
compounds, decomposition of 25 did not lead to nitrogen offꢀgassing. The inherent reactivity
of 25 instead resulted in electrocylization to form pyrazole 28 (Figure 4). In the solid state,
11
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 44
1
2
3
4
5
6
7
8
9
this electrocyclization event had an onset temperature of 42 °C, and a heat of reaction of 872
J/g. In solution, it was shown that this reaction was solvent and concentration dependent.14
Calorimetry studies revealed that solutions of ≤10 wt% of 25 in nꢀheptane were suitable for
processing, wherein the rate of the decomposition to 28 was approximately 3ꢀ5 % per day.
One additional advantage of this electrocyclization decomposition mode was that the resulting
pyrazole 28 was not soluble in nꢀheptane, and could easily be filtered off before the solution
was introduced to the cyclopropanation reaction. Potency of the nꢀheptane solution of 25
would slowly decrease over time, however the purity of the reagent solution was always
maintained.15
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 4. Electrocyclization of 25 to Afford Pyrazole 28
The styrene substrate 26 was generated in a twoꢀstep sequence (Scheme 6). In the first
step the commercially available 2ꢀmethoxybenzaldehye (29) was brominated in glacial AcOH
at 30–35 °C. Addition of water led to the direct precipitation of the crude product (30). The
resulting wet cake was reslurried with a dilute sodium bicarbonate solution to remove any
residual acid, affording the desired pure 2ꢀbromoꢀ5ꢀmethoxybenzaldehyde 30 in 83–90%
isolated yield.16 Selection of the appropriate solvent, AcOH, was critical to minimize overꢀ
bromination. The aldehyde 30 was then subjected to a Wittig olefination in THF. nꢀHeptane
was added followed by an aqueous workup. Optimization of the reaction and workup
conditions led to a facile crystallization of the triphenylphosphine oxide reaction byꢀproduct.
However, the cyclopropanation reaction proved highly sensitive even to residual levels of
triphenylphosphine oxide, so the mixture was passed through a basic alumina plug.
Surprising, we still observed variable catalyst performance. Careful evaluation of the mixture
revealed the culprit as very low levels of triphenylphosphine (200 ꢀ 400 ppm). To address the
trace levels of this highly impactful impurity, an oxidative wash with sodium percarbonate
and hydrogen peroxide was implemented during the aqueous workup. This converted any
residual triphenylphosphine to triphenylphosphine oxide, which was readily removed via the
original removal strategy (crystallization and passage through a basic alumina plug). After
12
ACS Paragon Plus Environment

Page 13 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
distillative solvent reduction, pure styrene 26 was obtained as a solution in nꢀheptane.17 Two
batches were prepared on kilogram scale. The first was produced as a 31.2 wt% nꢀheptane
solution in 64.7% yield and the second batch was prepared as a 21.5 wt% nꢀheptane solution
in 54.8% yield from aldehyde 30.
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 6. Synthesis of Styrene 26.
With access to nꢀheptane streams of the styrene 26 as well as the vinyl diazoester 25,
the cyclopropanation chemistry was investigated. The reactor was charged with 1.1
equivalents of 26 relative to diazoester 25 (the limiting agent). The Rh2(SꢀDOSP)4 catalyst
was charged, and the diazoester 25 was added over approximately 1 hour at room temperature.
Lab experiments demonstrated that 0.2 mol% catalyst loading could afford a 90% solution
yield and 85–86% ee. On scaleꢀup, enantioselectivity ranged between 80.6 and 83.2% ee.
Upon reaction completion nꢀheptane was removed via distillation and methylene chloride was
charged.
Scheme 7. Rhodium Catalyzed Asymmetric Cyclopropanation of Styrene 26.
13
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The crude reaction stream, which consisted of a ~ 10:1 mixture of cyclopropane 24 to
residual styrene 26, was then subjected to several telescoped functional group transformations
to afford crystalline 23•(R)-AMBA (Scheme 8). This approach was necessary both due to the
lack of crystalline intermediates and the desire to avoid diminishing the ee during isolations.
Scheme 8. Telescoped Functional Group Transformations of Vinyl Cyclopropane 24
The DCM solution of cyclopropane 24 was first cooled to ꢀ60 ˚C and exposed to a
stream of ozonized air while maintaining the internal temperature ≤ ꢀ40 ˚C. Upon complete
consumption of 24 a methanolic solution of sodium borohydride was then introduced to
14
ACS Paragon Plus Environment

Page 15 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
reduce the ozonide to the corresponding alcohol 31.18 The mixture was quenched with an
aqueous HCl solution and an extractive workup afforded a DCM solution of 31. Along with
desired product 31, the expected benzyl alcohol byꢀproduct 32 and alcohol 33, resulting from
ozonolysis and reduction of the excess 0.1 equivalents of the styrene 26, were produced
(Figure 5). Solvent exchange to MeOH/THF and subsequent introduction of an aqueous
LiOH solution effected ester hydrolysis to afford a solution of 23•Li. Extraction of the
reaction mixture with MTBE served to conveniently remove the two benzylic alcohols 32 and
33, along with other organic impurities, affording a relatively pure aqueous solution of lithium
carboxylate 23•Li.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Alcohol Impurities Generated after Ozonolysis and Reduction
The aqueous lithium carboxylate stream was then acidified, and the corresponding
acid was extracted into dichloromethane. While the free acid is crystalline (mp > 140 °C), we
found that isolation from a variety of solvents failed to improve the optical purity. The
dichloromethane solution was solvent exchanged to IPA:water, warmed to 60 °C, and treated
with (R)ꢀAMBA resulting in the crystallization of 23•(R)-AMBA. This process performed
well in the lab, furnishing approximately 65% isolated yield of optically enriched 23•(R)-
AMBA. At pilot scale, this telescoped sequence afforded 61 kg with a 53% isolated yield in
high chemical and optical purity (>99.0 HPLC area percent purity, >98 % ee).
To our knowledge, this was the first use of the Davies catalyst system on multiꢀ
kilogram scale. It was ideal from a fitꢀforꢀpurpose perspective as it identified a key
intermediate to control quality and facilitate isolation (23•(R)-AMBA), was developed very
quickly, and successfully enabled the delivery of early clinical supplies. The overall route to
15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 44
1
2
3
4
23•(R)-AMBA was 7 linear steps, but only 2 isolations, affording high quality product in an
5
6
overall 26% yield (Scheme 9).
7
8
9
Scheme 9. First Generation Synthesis of the Chiral Cyclopropane 23.
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
It was recognized there would be a number of challenges for implementation of this 1st
generation cyclopropanation chemistry in the commercial environment. The issues identified
to be addressed or mitigated with the next generation approach19 included: (1) Cost of the
chiral catalyst; (2) Stability of the diazo compound and safety concerns; (3) The sensitivity of
the catalyst to impurities in the substrates or reagents which could result in lower than desired
ee’s or conversions; (4) Difficulty directly isolating material in high ee and yield from a
reaction stream which was only 82ꢀ86% ee; (5) Specialized equipment and facilities to
accommodate the diazo chemistry as well as the ozonolysis chemistry.
Conversion of the Chiral Cyclopropane to an Alkylating agent
As mentioned previously, 23•(R)-AMBA was targeted because it could be readily
converted to the desired cyclopropyl tosylate 17 through simple functional group conversions
(Scheme 4). In our initial investigations, we were freebasing the salt and removing the (R)ꢀ
AMBA, however we quickly found that this extra step was not required. Esterification could
be done directly via the in-situ generation of anhydrous HCl with trimethylsilylchloride
16
ACS Paragon Plus Environment

Page 17 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
(TMSCl) in EtOH. One of the challenges with the esterification is related to the fact that 23
has alcohol functionality and can selfꢀesterify. Although this dimer was not impactful to
quality as the ester is ultimately hydrolyzed, high levels would correspond to decreased yields,
so two control points (concentration and temperature) were identified. The concentration of 9
L/kg was a good balance between minimizing dimer formation observed under high
concentration reaction conditions as well as managing volumes for efficiency, throughput, and
cost. Upon reaction completion, the bulk of the ethanol was removed via distillation, again to
control volumes. By operating under vacuum, it was possible to maintain the temperature at
≤40 °C, thus minimizing further dimerization. After an aqueous workup which served to
remove the one equivalent of (R)ꢀAMBA, the final toluene solution was azeotropically
distilled. Development of the tosylation reaction quickly identified DABCO and tosyl
chloride (TsCl) as the ideal reagents. The DABCO not only served as a base, but it also
activated the TsCl and promoted fast reactions (1–2 h).20 The toluene stream was treated with
DABCO and TsCl to generate the desired 17 in high inꢀprocess yield (> 90%). Levels of the
dimer were typically at 0.6ꢀ1.1 HPLC area percent. After an aqueous extraction, the mixture
was azeotropically distilled under vacuum and the product was crystallized by charging
heptane resulting in good purging of the dimer (<0.3 HPLC area percent).
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 highly crystalline cyclopropyl tosylate 17 was utilized for our development
activities on the subsequent alkylation reaction. However, a safety concern emerged with 17
prior to scaleꢀup. Differential scanning calorimetry (DSC) experiments at variable ramp rates
showed shifting degradation temperatures indicating 17 was susceptible to autocatalytic
decomposition. We theorized decomposition occurs via acid catalyzed cation formation and
subsequent ring opening of the cyclopropane. This theory was supported when it was
demonstrated small amounts of pꢀtoluenesulfonic acid could trigger this very exothermic
decomposition (ꢀ243 J/g, Tadiabatic 146 ºC).
It was determined that a 4 L/kg solution of 17 in toluene could be held for several
days, even at temperatures above the conditions required for the alkylation chemistry to occur.
More importantly, if uncontrolled decomposition was to occur, the toluene stream would not
pose a safety risk (ꢀ30 J/g, Tadiabatic 17 ºC). The inherent stability liability of isolated 17
required us to rapidly adjust our strategy and develop a telescoped process which did not
require the isolation, storage and handling of 17 as a solid.
17
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 44
1
2
3
4
Alkylation of Indole 16
5
6
7
8
With access to the indole 16 and cyclopropyl tosylate 17 fragments, we set out to
develop the alkylation chemistry to couple the two fragments. Identifying the optimal
reaction conditions proved to be challenging. A solvent and strong base system that could
generate the relatively insoluble indole dianion and facilitate the alkylation chemistry was
required. NaHMDS in DMF was selected after initial screening. While the alkylation product
15 could now be formed in good inꢀprocess yields, the crystallization also proved to be highly
challenging. It was found that while pure 15 could be crystallized from THF/IPA, the
crystallization was simply not robust as it was highly sensitive to impurities (including
unreacted starting materials 16 and 17) which resulted in low isolated yields and product
purity.
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
The most significant challenge with the alkylation chemistry was the sensitivity to
reagent and base stoichiometry, especially given the lack of a robust crystallization. It was
critical to maintain a slight excess of indole 16 relative to cyclopropyl tosylate 17.
Development work identified a narrow range of 1.05ꢀ1.10 equivalents of indole 16 relative to
17 under which high conversion could be obtained with minimal remaining indole 16. In
addition, it was also critical to utilize slightly less than 2.0 equivalents of base relative to
indole 16, with the desired range identified as 1.90ꢀ1.95 equivalents. If the base was
undercharged, lower inꢀprocess yields were observed and much lower isolated yields were
obtained. If the base was overcharged, epimerization of the cyclopropane to compound 34
was observed (Figure 6). Even low levels of 34 resulted in dramatic losses of product to the
mother liquor during the isolation.
Figure 6. Alkylation Epimerization Impurity 34
18
ACS Paragon Plus Environment

Page 19 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The requirement of utilizing cyclopropyl tosylate 17 as a toluene solution, instead of a
crystalline intermediate, added extra complexity to a reaction and crystallization already
highly sensitive to stoichiometry. The approach adopted was to generate the toluene stream of
17 and then charge 90% of the theoretical quantity of indole 16. As the exact quantity of 16
was known, it could be used as an internal standard and HPLC analysis of the mixture then
enabled calculation of the potency of the cyclopropyl tosylate reaction stream as well as the
appropriate quantity of additional 16 to charge. Although the base was titrated prior to use,
we took similar precautions with the base charge due to the severe consequences to yield that
would result from epimerizing the stereocenter. A 90% charge of base, sampling and HPLC
analysis enabled accurate calculation of the additional base charge required.
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
This overall approach (Scheme 10) was fitꢀforꢀpurpose for the initial kilogram scale
campaign. In a 300 g demonstration run, we achieved 74% yield and 99.0 HPLC area percent
purity. We charted a conservative approach on scale and deliberately avoided charging
additional base to reach full conversion of cyclopropyl tosylate 17, accepting a moderate loss
of product to mother liquor versus a potential catastrophic loss of product to epimerization.
This chemistry was used to produce 16 kg of 15 in 98.6 HPLC area percent and 46.7% yield
in one batch for the initial scaleꢀup. This chemistry was repeated to produce 181 kg of 15 in ≥
92 HPLC area percent and 39, 51, and 61% yield over three sequential batches. The lack of
robustness in the crystallization and familiarization requirement to approach the maximum
potential yields in the pilot plant setting was a clear signal that this chemistry would need
additional development before being considered acceptable as a commercial route.
Scheme 10. Tosylate 17 Generation and Alkylation of Indole 16
19
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Intramolecular Pd Catalyzed Direct Indole Arylation
Initial work on the cyclization chemistry was conducted on a model substrate 35
(Scheme 11). Screening of solvents showed DMF and DMAc were comparable while NMP
afforded increased levels of a desꢀbromo impurity 37. It was also found that control of water
was important as increased water content afforded higher levels of 37. A screen of catalysts
was conducted with initial results suggesting impurity 37 levels increased with Pd(dba)2 and
Pd(tfa)2 catalyst systems. Pd(OAc)2 proved to be superior with loadings as low as 5 mol%
being able to effect full conversion. A series of ligands were evaluated as well (Figure 7) with
all proving to be inferior to PCy3HBF4. Interestingly, carbene ligand 40 afforded large
amounts of the impurity 38 which likely results from a Pd migration to the cyclopropyl ring.21
The initial proof of concept was conducted with potassium carbonate (K2CO3) as the base,
however potassium acetate (KOAC) was chosen for the initial scaleꢀup as less desꢀbromo
impurity 37 was observed. Optimal conditions were defined as using 5 mol % Pd(OAc)2, 10
mol % PCy3•HBF4, and 5 equivalents of KOAc affording (inꢀprocess) 72 HPLC area percent
of product 36 and 20 HPLC area percent of 39, corresponding to the hydrolysis of the methyl
ester adjacent to the cyclopropane ring.
Scheme 11. Development of Pd Catalyzed Direct Arylation on Model Substrate 35
20
ACS Paragon Plus Environment

Page 21 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Ligands Evaluated for Pd Catalyzed Direct Arylation
With the optimal conditions in hand from the model substrate, compound 15 was
subjected to the same conditions (Scheme 12). The ethyl ester protecting group provided
increased hydrolytic stability versus the methyl ester used in the model system, thus
minimizing the main hydrolysis impurity 13. Even though the hydrolysis impurity is actually
the next intermediate and poses no risk to quality, removal during the workup of ester 41
results in decreased yields, and thus was to be avoided. We were pleased to see the conditions
from the model system translated well to the actual substrate and proved to be relatively
robust chemistry. The reaction could be stopped by cooling and restarted by heating back to
the reaction temperature (120 °C). A catalyst reload could drive the reaction to completion.
21
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 44
1
2
3
4
5
6
7
8
9
Upon reaction completion, the DMAc was distilled, MTBE was charged and extractions with
aqueous solutions of citric acid afforded an MTBE solution of the cyclized product 41.
Treatment with an aqueous KOH solution required high agitation to effect the hydrolysis as it
was a biphasic solution. After the hydrolysis was complete, citric acid neutralization and
extractions removed the phosphine ligand and afforded an MTBE solution of the hydrolyzed
product 13. A solvent swap to EtOH and treatment with a 24% KOEt solution enabled
crystallization and isolation of product 13 as the potassium salt. The solid was determined to
be an ethanol solvate (containing 7ꢀ9% ethanol) (13•Κ•EtOH). Drying at temperatures
above 50 °C led to discoloration so thermal desolvation was not an option. A second crop
was then obtained from the mother liquor. The major impurity observed was the desꢀ
bromination impurity 42 (Figure 8) which increases with water content. This process was
used to produce 9.14 kg in an average 68.2 % yield with 98.6 ꢀ 99.8 HPLC area percent purity
and ~110 ppm Pd over three batches. While effective, this process had a number of
drawbacks. This included the high vacuum distillation of DMAc, twoꢀcrop isolation, residual
palladium levels, long reaction times (12ꢀ14 h) and a tedious workup procedure.
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 Pd Catalyzed Intramolecular Direct Indole Arylation
22
ACS Paragon Plus Environment

Page 23 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Des-bromo Impurity 42
A number of modifications were made to improve the overall process prior to the next
scaleꢀup. An extensive screen of bases (KOtBu, K2CO3, KOAc, Cs2CO3, CsOAc, CsHCO3,
Na2CO3, NaHCO3, K3PO4, K2HPO4, KH2PO4, AgOAc, Ag2CO3, DBU, and triethylamine) led
to the selection of KHCO3 which decreased reaction times from 12ꢀ16 h to ~4 h. The reaction
solvent was changed from DMAc to a 1:1 DMAc:toluene mixture. After cyclization is
complete, water was charged followed by solid KOH to perform ester hydrolysis. A filtration
was performed to remove palladium black that has precipitated out, and the toluene layer
containing desꢀbromo impurity 42 was discarded. After neutralizing the aqueous phase with
HCl, the desired product was simply extracted into MTBE obviating the need for a DMAc
high vacuum distillation. It was also found that the citric acid washes could be removed. In
addition, a small reꢀcharge of DMAc prior to the crystallization afforded a hemi DMAc
solvate which afforded the desired product in a high yielding 1ꢀdrop crystallization and
avoided the formation of an ethanol solvate. This process (Scheme 13) was used to produce
139 kg of 13•Κ•1/2DMAc as the potassium salt/hemi DMAc solvate in 85% yield with >99.6
HPLC area percent purity and < 100 ppm Pd.
Scheme 13. Modified Pd Catalyzed Intramolecular Direct Arylation
23
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
BMS-791325 (1) Development and Synthesis
Our investigation into the API (active pharmaceutical ingredient) step began with an
evaluation of the Discovery conditions which employed 2ꢀ(1Hꢀbenzotriazoleꢀ1ꢀyl)ꢀ1,1,3,3ꢀ
tetramethylaminium tetrafluoroborate (TBTU). We quickly found the TBTU chemistry was
not amenable for scaleꢀup as TBTU activation was accompanied by the formation of a number
of impurities and the reaction mixtures were typically very highly colored (dark redꢀbrown).
Similar to our Discovery colleagues, we were also forced to employ preparative HPLC
separation in order to obtain material of the requisite purity. While evaluating alternative
coupling
reagents,
we
rapidly
narrowed
our
focus
to
1ꢀethylꢀ3ꢀ(3ꢀ
dimethylaminopropyl)carbodiimide (EDAC), 1ꢀhydroxybenzotriazole (HOBt) and N,Nꢀ
diisopropylethylamine (DIPEA) conditions as there was extensive experience with many
successful onꢀscale implementations of this chemistry.22 We first sought to determine
whether the EDAC/HOBt protocol was viable in terms of reactivity (note the sterically
congested nature of the amide bond that is formed in this process) and impurity profile. We
were pleased to find that the EDAC/HOBt chemistry (Scheme 14) was very wellꢀsuited for
the preparation of BMS-791325 (1) via coupling of compound 13•Κ•EtOH and the bridged
piperazine 14.
Scheme 14. EDAC and HOBt Coupling of 13•K and 14 to Produce BMS-791325 (1)
24
ACS Paragon Plus Environment

Page 25 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Solvents that were assessed for the reaction were ACN, THF, DMF and 2ꢀ
methyltetrahydrofuran (2ꢀMeTHF). ACN and THF were found to behave equally well, with
both giving excellent HPLC purity profiles for the reaction at completion. DMF was
satisfactory as a reaction solvent, but had the potential for prohibitively high volumes during
the workup. 2ꢀMeTHF did not perform well; specifically, the reaction proceeded at a slower
rate when compared with acetonitrile and THF, presumably due to diminished solubility of
the reaction components in 2ꢀMeTHF. As discussed (vide supra), our initial development
work and scaleꢀup to produce intermediate 13 would inevitably contain high levels of water
and EtOH (~10% by weight). In the API step, EtOH was shown to generate the ethyl ester
impurity, and excess water caused reaction stalling. As a result, the API step required
modification to contend with these contaminants. Azeotropic removal of EtOH and water was
found to be an effective solution, and in this regard, acetonitrile proved to be far superior to
THF in its ability to purge EtOH and water via azeotropic distillation. Thus, acetonitrile was
selected as the reaction solvent.
The primary impurities of concern in the amidation chemistry are the Nꢀacylureas
(compounds 43 and 44, Figure 9) that form via isomerization of the primary Oꢀacylurea.23
These compounds are dead ends on the reaction pathway. They form in approximately a 2:1
ratio favoring isomer 44. Implementation of HOBt, at loadings of > 0.5 equivalents,
minimized the formation of the Nꢀacylurea impurities 43 and 44. From the perspective of
25
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 44
1
2
3
4
overall reaction purity/impurity profile and reaction rate (our targeted time to reaction
5
6
completion was ≤ 10 hours), 1.15 equivalents of HOBt was found to be an optimal loading.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Observed N-Acylurea Impurities 43 and 44
The other components in this reaction (EDAC, DIPEA, and the bridged piperazine 34)
were not subjected to extensive rangeꢀfinding studies at the time of initial scaleꢀup. EDAC
and 14 loadings were simply adjusted to match the HOBt loading, and DIPEA was adjusted to
approximate the stoichiometric HCl loading in the reaction system. In the reaction sequence,
reagents were added in the following order: 14, HOBt, EDAC and DIPEA and delays of up to
1 hour in the addition of each reagent were proactively explored and found to be well
tolerated.
The workup is designed to minimize losses to the aqueous layers via maintaining
aqueous pH within the isoelectric zone (~pH 4.5 to 7.0) for the API as the aqueous solubility
of BMSꢀ791325 increases dramatically outside of this range. For the reaction quench, a half
saturated aqueous ammonium chloride solution was found to effectively extract residual
EDAC, the corresponding EDACꢀderived urea (EDCU), DIPEA and residual 14 into the
aqueous layer. However, quenching the reaction directly into this solution led to significant
loss of API due to the fact that the resulting aqueous pH is approximately 10. To solve this
problem, five equivalents of acetic acid were added to the quench medium. As a result, the
postꢀquench aqueous pH was typically about 5, a value that was within the desired isoelectric
range. Implementation of the acetic acid charge had an additional benefit in that the aqueous
pH does not fall significantly below the lower limit of the isoelectric range, due to the pKa of
this acid. The phase split at pH~5 still removed most of the excess bridged piperadine 14,
DIPEAꢀHCl, EDCU and residual EDAC.
26
ACS Paragon Plus Environment

Page 27 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The derived organic layer still contained HOBt. In most amidation reactions, HOBt is
removed by extraction with an aqueous base.24 In our case, this was not a viable approach due
to the aqueous solubility of BMS-791325 at pH values greater than 7. As a result, we chose
to extract the organic layer with pH 7 phosphate buffer (K2HPO4/KH2PO4), an aqueous
medium that provided good extraction efficiency for HOBt while maintaining the aqueous pH
in the desired range. It is important to note that the choice of coꢀsolvent for the workup is
critical to the effectiveness of the phosphate buffer extraction. It was found that
isopropylacetate (IPAc) provided 2x better partitioning of HOBt into the aqueous layer when
compared to ethyl acetate, and 10x better partitioning when compared to 2ꢀMeTHF. Two
extractions with 12.5 volumes of phosphate buffer led to consistent removal of HOBt to ≤ 2
HPLC area percent relative to BMSꢀ791325.
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
Following the workup sequence, the IPAc:ACN mixture is subjected to constantꢀ
volume distillation to effect solvent exchange to IPAc. Laboratory studies showed that ACN
is typically reduced to nonꢀdetect levels25 prior to meeting a KF specification of ≤500 ppm.
Therefore, we only monitored KF as a solvent exchange endpoint thus reducing the analytical
burden. The KF specification of ≤500 ppm maximized the effectiveness of the upcoming
clarification to remove residual inorganic impurities and other insolubles. A low water level
at this stage was also desired because a favorable ethanol:water azeotrope is not accessible.
Due to limited solubility of BMSꢀ791325 in IPAc, we elected to solvent exchange to ethanol
prior to clarification via a polish filtration. Once the solution was filtered, a 1.25N solution of
HCl in ethanol was slowly charged, the batch was seeded and MTBE was charged over 1 h.
The batch was held at 25 °C for 12 h, the resulting slurry of the BMS-791325•HCl salt was
filtered and the cake was washed with an MTBE:ethanol mixture. This process, in four
batches, afforded a total of 6.39 kg of 1 in 76.4% average yield26 in > 99.97 HPLC area
percent purity and > 99.9 ee in 4 batches. A subsequent delivery afforded 68.9 kg in 78.5%
average yield27 in >99.6 HPLC area percent purity and >99.9 % ee in 3 batches.
One of the challenges with the crystallization of BMS-791325•HCl is the fact that up
to 16% of the product often remained in the mother liquors postꢀcrystallization despite all
efforts to optimize antiꢀsolvent ratio and temperature of the crystallization. This was
identified as something that would have to be addressed in subsequent development work.
27
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 44
1
2
3
4
5
6
7
8
Conclusions
In summary, we describe an efficient, highly convergent synthesis (Scheme 15) of
BMS-791325 HCl (1) which obviated the need for an end game chiral resolution. The chiral
cyclopropane fragment (23 (R)-AMBA) was prepared in a highly telescoped fashion (7 linear
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
•
steps and only 2 isolations) in 26 % yield via asymmetric catalysis using the Rh2(SꢀDOSP)4
catalyst. This is the first known kilogramꢀscale demonstration of the Davies methodology.
The indole fragment was prepared in 6 steps and 3 isolations in 53% yield. After the
cyclopropane fragment is activated and indole is Nꢀalkylated, the palladium catalyzed
cyclization provides quick and direct access to the 7ꢀmembered ring. Hydrolysis and coupling
affords BMS-791325•HCl (1) in high quality. The telescoped endgame is conducted in 6
steps and 3 isolations in 32% yield. The overall process has been demonstrated on multiꢀ
kilogram scale and was considered fitꢀforꢀpurpose successfully fulfilling initial toxicological
and clinical requirements and laying the ground work for development of the commercial
process. The target was prepared in 12 linear steps with 5 isolations in an overall yield of
8.2%.
Scheme 15. Complete Linear Synthesis of BMS-791325•HCl (1)
28
ACS Paragon Plus Environment

Page 29 of 44
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Experimental Section
All reactions were performed under a nitrogen atmosphere unless otherwise noted.
Reagents were used as received from vendors, unless otherwise noted. All reagents purchased
from vendors were used as received unless otherwise indicated. NMR spectra were recorded
on a Bruker DRXꢀ400 or AV III 600 instrument and are referenced to residual undeuterated
solvents. The following abbreviations are used to explain multiplicities: br = broad, s =
singlet, d = doublet, t = triplet, q = quartet, m = multiplet. Highꢀresolution mass spectrometry
(HRMS) was performed on an Agilent 6230B TOF instrument.
Preparation of methyl 3-cyclohexyl-1H-indole-6-carboxylate (3).
To a reactor was charged methyl indoleꢀ6ꢀcarboxylate (2) (85.0 kg, 485.2 mol, 1.0
o
equiv.) and toluene (206.0 kg). The resulting slurry was stirred for 15 min at 20 C. To this
mixture was charged cyclohexanone (94.0 kg, 957.7 mol, 2.0 equiv.) while maintaining the
o
temperature at 20 C. To the solution was charged Et3SiH (172.0 L, 1479.2 mol, 3.0 equiv.)
in 3 portions. The mixture was cooled to 0 oC. Trifluoroacetic acid (172.0 kg, 1508.5 mol,
3.1 equiv.) was charged over 90 minutes controlling the temperature below 30 °C (note: this
addition is exothermic and requires active cooling). After the addition was complete, the
29
ACS Paragon Plus Environment