Development of a Synthesis of a 2,3-Disubstituted 4,7-Diazaindole Including Large-Scale Application of CH3Li/TiCl4-Mediated Methylation of an Enolizable Ketone

Article abstract of DOI:10.1021/op5003769

The chemical development of a 2,3-disubstituted 4,7-diazaindole is described. The requisite tertiary carbinol substrate was prepared employing in situ-generated CH₃TiCl₃ as a chemoselective and preferred reagent compared to CH₃MgX for methyl addition to an enolizable ketone. The 4,7-diazaindole ring system was efficiently assembled via an intramolecular Chichibabin transformation. The optimized processes were performed on pilot-plant scale to provide kilogram quantities of the target molecule.

Full text of DOI:10.1021/op5003769

Subscriber access provided by NEW YORK UNIV
Full Paper
Development of a Synthesis of a 2,3-Disubstituted-4,7-
Diazaindole Including Large-Scale Application of CH3Li/
TiCl4-Mediated Methylation of an Enolizable Ketone
Franz J Weiberth, Harpal S. Gill, George E. Lee, Duc S. Ngo, Frederick L. Shrimp, Xuemin Chen, Geoffrey
D'Netto, Bryan R Jackson, Ying Jiang, Narendra Kumar, Frederick Roberts, and Evgeny Zlotnikov
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/op5003769 • Publication Date (Web): 02 Jun 2015
Downloaded from http://pubs.acs.org on June 16, 2015
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 free 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 accessible to all
readers and 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.
Organic Process Research & Development 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 15
Organic Process Research & Development
1
2
3
4
Development of a Synthesis of a 2,3-Disubstituted-4,7-Diazaindole Including Large-
Scale Application of CH3Li/TiCl4-Mediated Methylation of an Enolizable Ketone
5
6
7
8
9
Franz J. Weiberth,* Harpal S. Gill,# George E. Lee,# Duc P. Ngo,# Frederick L. Shrimp II,#
Xuemin Chen,# Geoffrey D’Netto,# Bryan R. Jackson,# Ying Jiang,# Narendra Kumar,^
Frederick Roberts^ and Evgeny Zlotnikov#
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 Development, Sanofi U.S. R&D, 153 Second Ave, Waltham, Massachusetts
02451, U.S.A.
Abstract:
The chemical development of a 2,3ꢀdisubstitutedꢀ4,7ꢀdiazaindole is described. The
requisite tertiary carbinol substrate was prepared employing in situꢀgenerated CH3TiCl3 as
a chemoselective and preferred reagent compared to CH3MgX for methyl addition to an
enolizable ketone. The 4,7ꢀdiazaindole ring system was efficiently assembled via an
intramolecular Chichibabin transformation. The optimized processes were performed on
pilotꢀplant scale to provide kilogram quantities of the target molecule.
Introduction
The original synthesis of 1 (Scheme 1), a compound under development as a Syk
inhibitor,1 suffered from several limitations. For example, 2ꢀpropylpyrazine (4) was
prepared by alkylation of 2 with ethyl iodide under cryogenic reaction conditions. In
addition, the isolation of 4 required a tedious fractional distillation in order to remove
* Author to whom correspondences should be sent via email: franz.weiberth@sanofi.com
# Chemical Development, Sanofi U.S. R&D, Bridgewater, NJ, at the time this work was conducted.
^ Chemical Development Support, Sanofi U.S. R&D, Bridgewater, NJ, at the time this work was
conducted.
1 Gillespy, T. A.; Eynott, P.; Allen, E. M.; Yu, K. T.; Zilberstein, A. PCT Int. Appl. US
2010/0035884 A1, 2010.
1 of 15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 15
1
2
3
4
5
6
7
8
9
unreacted 2 and some 2ꢀ(pentanꢀ3ꢀyl)pyrazine side product that resulted from diethylation,
both of which subsequently generated related impurities in downstream steps that were
difficult to remove. Also, the reaction of the anion of 4 with the nitrile 5 and subsequent
cyclization provided 6 in only about 25% yield. The overall yield of 1 was only about
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
12%.
Consequently, in order to support the development program with multiꢀkg
quantities of 1, a more efficient and scalable synthesis was required.
Scheme 1. Original Synthesis of 1
O
N
N
5
N
LDA / THF
+
I
- 60 oC
82%
NaN(Si[CH]3)2
N
N
THF, dioxane, 0 oC
25-30%
4
3
2
1. CH3MgCl, THF, 0oC
N
N
N
N
O
N
N
OH
2. recryst.
60-70%
H
H
6
1
Results and Discussion
It was envisioned that the issues in the original route related to preparation of 4 could be
mitigated by crossꢀcoupling nꢀPrMgX with 2ꢀchloropyrazine as described by Fürstner.2
An initial probe of the ironꢀcatalyzed crossꢀcoupling method demonstrated feasibility and
superiority over the original ethylation process, and was adapted to prepare early supplies
of 2ꢀpropylpyrazine. The protocol consisted of adding 1.2 equiv of 1ꢀpropylmagnesium
2 (a) Fürstner, A.; Leitner, A.; Méndez, M.; Krause, H. J. Am. Chem. Soc. 2002, 124, 13856. (b) Scheiper,
B.; Bonnekessel, M.; Krause, H.; Fürstner, A. J. Org. Chem. 2004, 69, 3943.
2 of 15
ACS Paragon Plus Environment

Page 3 of 15
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
chloride to 2ꢀchloropyrazine (7, Scheme 2) in the presence of 5 mol% Fe(acac)3 at 10ꢀ25
oC over a 2ꢀh period. Best conversions, achieving >97 area % (A%3) product, were
obtained using THF as solvent compared to MTBE (77%), NMP/THF (4:1; 55%) and
NMP:MTBE (1:1; 94%). The reaction was typically complete after stirring for an
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
additional 10 min at 20 C. Minor amounts of two side products were observed at <2%
each, namely, pyrazine from dechlorination of the starting material and bipyrazine from a
homocoupling pathway. After an aqueous quench, an extractive workup that included a
wash with aqueous NaOH to remove 2,4ꢀpentanedione derived from the catalyst,
concentration, and simple distillation, 4 was obtained in 62ꢀ77% yield and >99.5% purity.
This process was employed to prepare 4 used in preꢀclinical manufacturing campaigns of
1. As the project progressed, a bulk supplier emerged that was able to provide 4 from
stock in >10ꢀkg quantities and this material was utilized in subsequent manufacturing
campaigns.
The reaction of the anion of 4 with a suitable aryl nitrile with subsequent intramolecular
Chichibabin cyclization4 was deemed to be a concise strategy for assembling the 4,7ꢀ
diazaindole core structure in the target molecule. Complications in the original route
(Scheme 1) were caused by enolate formation in two steps of the synthesis, namely, during
the Chichibabin step and the Grignard methylation that led to poor conversions and the
formation of aldol byproducts. An alternative approach that was envisioned to address
enolate formation was reversing the Chichibabin and Grignard methylation sequences by
methylating 4ꢀacetylbenzonitrile prior to Chichibabin cyclization. In this approach, the
3
All references to reaction conversion and purity are understood to be in terms of HPLC relative area %
unless indicated otherwise.
4 Vierfond, J.ꢀM.; Mettey, Y.; MascrierꢀDemagny, L.; Miocque, M. Tetrahedron Lett. 1981, 22, 1219.
3 of 15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 15
1
2
3
4
5
6
7
8
ketone functionality would be present in only one step of the synthesis, and complications
during the Chichibabin step from enolate formation of the methyl ketone would be
eliminated.
9
Indeed, early lab results demonstrated feasibility and potential scalability for both the
crossꢀcoupling reaction to prepare 4 and the modified Chichibabin protocol, and this
alternative, and now more convergent, route was selected for further development (Scheme
2).5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Convergent Synthesis of 1
N
Cl
N
n-PrMgCl
cat.
N
N
7
4
N
N
base
N
OH
H
N
N
CH3MgBr
1
O
HO
5
8
Tertiary carbinol 8 prepared using CH3MgX
Initial supplies of 4ꢀ(1ꢀhydroxyꢀ1ꢀmethylethyl)benzonitrile (8) were prepared by
methylation of 4ꢀacetylbenzonitrile with 1.25 equiv of methylmagnesium bromide (3 M in
diethyl ether) in methyl tertꢀbutyl ether (MTBE) (Scheme 3).6 Partial enolization of 4ꢀ
acetylbenzonitrile caused by the basic methylmagnesium bromide resulted in incomplete
consumption of starting material (2ꢀ15% remaining) and the formation of about 5ꢀ15% of
5 Lee, G. E.; Shrimp II, F. L.; Weiberth, F. J. PCT Int. Appl. WO 2010/107969 A1, 2010
6 Popielarz, R.; Arnold, D. R. J. Am. Chem. Soc. 1990, 112, 3068.
.
4 of 15
ACS Paragon Plus Environment

Page 5 of 15
Organic Process Research & Development
1
2
3
4
5
6
7
8
the aldol sideꢀproduct 9.7,8 The low solubility of the magnesium enolate of 5 in ether
solvents, especially MTBE, prevented further conversions to 8 and limited the amounts of
9. Similarly, the insolubility of the magnesium alkoxide of 8 contributed to achieving high
chemoselectivity by minimizing overꢀreaction of the Grignard with the nitrile functionality
to afford 10.
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
Scheme 3. Reaction of 5 with CH3MgBr
O
N
N
N
N
CH3MgBr
HO
O
+
+
HO
OH
O
5
9
10
8
Typically, after 0.5 h at 10ꢀ15 °C following addition of Grignard reagent, the reaction
slurry contained a 68:18:13 ratio of desired product : starting material : aldol product
(compounds 8:5:9) as determined by HPLC analysis. The addition of THF (3 V)
facilitated some solubilization of the magnesium salts and further conversion of unreacted
starting material to achieve a 79:8:13 ratio of 8:5:9. After quench and an extractive
workup, the resulting MTBE phase was then partially concentrated and filtered to remove
some 9 that had crystallized to afford crude product as a concentrate containing a 89:5:5
ratio of 8:5:9. Distillation through a short path column (1ꢀ2 mbar, 125ꢀ128 °C) followed
by trituration with heptane, provided 8 as a white solid in 74ꢀ84% yield and >99% purity.
Although this preparation provided 8 in multiꢀkg quantities and in quality suitable for
downstream chemistry, the protocol was tedious, especially the purification. The
distillation of 8 to remove 5 in the forerun and 9 in the residue was not robust due to
thermal dehydration of 8 to concomitantly form the corresponding styrene in small
amounts. In some batches, perhaps due to variations in the quality of crude 8, retroꢀaldol
7 1H NMR (300 MHz, CHCl3, δ): 8.01 (d, J = 8.6 Hz, 2H), 7.79 (d, J = 8.6 Hz, 2H), 7.66ꢀ7.59 (m, 4H), 3.76
(d, J = 17.2 Hz, 1H), 3.52 (s, OH), 3.43 (d, J = 17.2 Hz, 1H), 1.65 (s, 3H).
8 (a) Batchelor, K. J.; Bowman, W. R.; Davies, R. V.; Hockley, M. H.; Wilkins, D. J. J. Chem Res. 1999,
428. (b) Tucker, J. L.; Couturier, M.; Leeman, K. R.; Hinderaker, M. P.; Andresen, B. M. Org. Process Res.
Dev. 2003, 7, 929.
5 of 15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 15
1
2
3
4
5
6
7
8
9
reaction of 9 was observed that contaminated fractions of 8 with 5 during the distillation.
Screening experiments were performed that evaluated several reaction parameters,
including reaction solvent, CH3MgCl vs. CH3MgBr, order of addition, and the temperature
and duration of reagent addition. These studies did not identify conditions that could
provide substantially improved conversions to 8. The use of 10 mol% ZnCl2 to suppress
enolate formation9 was not successful (>50% 9). Adapting a report from Knochel10 that
described the beneficial effects of soluble lanthanide salts towards favoring the addition
reaction over enolization, and using stoichiometric amounts of commercially available
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
LaCl3·LiCl in THF with CH3MgX, led to modest suppression of the aldol side product (5ꢀ
10%). Alternative substrates were also considered. An oxidative protocol using pꢀ
cuminonitrile11 was low yielding, and the aldol side reaction was still problematic when
ester substrates8 were employed.
Tertiary carbinol 8 prepared using CH3TiCl3
A search to identify a preferred methylating reagent that was a good nucleophile and a
12
poor base as a means to retard or eliminate enolization led to an evaluation of CH3TiCl3,
previously described by Reetz.13 Indeed, CH3TiCl3, prepared in situ from TiCl4 and
CH3Li, was found to react chemoselectively with 5 without enolization to provide 8 as the
sole product under optimized conditions (Scheme 4).
An attempt to replace CH3Li with an analogous Grignard reagent was studied in several
solvent systems, including, CPME, Bu2O and iꢀPr2O. In general, the reaction of CH3TiCl3
9 (a) Hatano, M.; Suzuki, S.; Ishihara, K. J. Am. Chem. Soc. 2006, 128, 9998. (b) See also: Hatano, M.; Ito,
O.; Suzuki, S.; Ishihara, K. J. Org. Chem. 2010, 75, 5008.
10 Krasovskiy, A.; Kopp, F.; Knochel, P. Angew. Chem. Int. Ed. 2006, 45, 497.
11 Tucker, J. L.; Couturier, M.; Castaldi, M. J.; Chiu, C. K. F.; Gestmann, D. Synth. Commun. 2006, 36, 2145.
12
For simplicity, this reagent is depicted as CH3TiCl3, although the actual species may be a solvate or
another complex depending on the specific reaction conditions.
13
Reetz, M. T.; Kyung, S. H.; Hüllmann, M. Tetrahedron 1986, 42, 2931ꢀ2935. See also: Reetz, M. T.
Organotitanium Chemistry. In Organometallics in Synthesis; Second Edition; Schlosser; M., Ed.; John
Wiley & Sons, Ltd.: England, 2002; Chapter VII, pp. 819ꢀ860.
6 of 15
ACS Paragon Plus Environment

Page 7 of 15
Organic Process Research & Development
1
2
3
4
5
6
produced from TiCl4/CH3MgBr 14 was less efficient than CH3TiCl3 produced from
TiCl4/CH3Li. Reaction times were longer (20 h vs. 2 h) and conversions were incomplete
(1ꢀ5% unreacted 5), even when 3 equiv of the methylating reagent were employed.15
7
8
o
Increases in the reaction time or temperature (25ꢀ30 C) were detrimental, leading to only
9
minor improvements in conversion while forming longꢀretention impurities.
Consequently, efforts were focused on using CH3TiCl3 produced from TiCl4/CH3Li rather
than TiCl4/CH3MgX based on better reaction performance and purity profile as the best
balance versus the handling of CH3Li on large scale.
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 4. Methylation using CH3TiCl3
N
5
CH3Li or CH3MgX
HO
CH3TiCl3
TiCl4
8
Several commercially available solution formulations of CH3Li were evaluated and were
found to be mainly equivalent in reaction performance, but with some differences in side
reactions. CH3TiCl3 prepared from CH3Li in ethyl ether (1.6 M) was not pursued beyond
lab batches because of restrictions on handling diethyl ether on scale. CH3TiCl3 prepared
from CH3Li in diethoxymethane (3 M) performed well in the methylation. But, upon
o
quenching of the reaction at ꢀ2 to 6 C, about 6% of the mixed acetal 11 formed (Scheme
5). Although this impurity could be avoided (<0.1%) by quenching at < ꢀ15 oC, there was
an additional concern that acid generated during workup may produce formaldehyde, a
genotoxin, from diethoxymethane. CH3TiCl3 prepared from CH3Li in cumene/2ꢀMeTHF16
also performed well in spite of the presence of two lithiated cumene impurities present at
levels <1 mol% in the commercial CH3Li solutions. These reactive impurities led to the
14
For an analogous application employing CeCl3/CH3MgCl, including pretreatment and activation of CeCl3,
see Larkin, J. P.; Wehrey, C.; Boffelli, P.; Lagraulet, H.; Lemaitre, G.; Nedelec, A,; Prat, D. Org. Process
Res. Dev. 2002, 6, 20, and related references cited therein.
15 Typically 1ꢀ5% unreacted 5 remained after more than 20 h at 20 oC.
16
Typically, 2.5ꢀ3.5 wt% CH3Li, 12ꢀ16 wt% 2ꢀMeTHF, <2 wt% heptane, with cumene making up the
balance.
7 of 15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 15
1
2
3
4
5
formation of about 1.5% of the corresponding addition products 12 and 13 when the
o
methylations were performed at 20 C (Scheme 6), but their levels were minimized
o
(<0.1%) by performing the methylations at ꢀ10 to 0 C. In case further mitigations were
6
7
8
9
needed, it was demonstrated that both 12 and 13 were readily purged into the filtrate
during isolation of 8. Reaction performance using CH3Li in cumene/THF was equivalent
to the cumene/2ꢀMeTHF formulation, including the generation and control of impurities 12
and 13, but there were supply chain issues and higher pricing for bulk quantities at the time
for this reagent formulation. As result, CH3Li in cumene/2ꢀMeTHF was selected as the
preferred reagent for further development.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Formation of mixed acetal 11
N
OH
CH2(OEt)2
+
N
O
O
8
11
.
Scheme 6. Addition products from metalated impurities in CH3Li
OH
OH
Li
Li
5
+
+
N
N
12
13
The effects of various ether solvents and other parameters, including stoichiometry,
reaction time and temperature, on methylation using CH3TiCl3 generated from equimolar
quantities of TiCl4 and CH3Li, were studied, and the results are summarized in Table 1.
Cryogenic temperatures were not necessary. The reagent was prepared by treating TiCl4
o
with CH3Li at 0 to ꢀ10 oC. Then, after aging for about 30 min at ꢀ5 C, 5 was added. The
aldol side reaction was observed when the methylation was performed in 2ꢀMeTHF, THF
and 1,2ꢀdimethoxyethane. The methylation was sluggish in MTBE. CPME and anisole
were the best solvents. About 1.5 equiv of CH3TiCl3 were needed for complete reaction of
o
5. The methylation was best performed at ꢀ10 to 0 C to avoid the formation of the
cumenyl side products observed when the reaction was performed at 20 oC (Entry 13).
8 of 15
ACS Paragon Plus Environment

Page 9 of 15
Organic Process Research & Development
1
2
3
4
5
Anisole was selected as the preferred solvent for the transformation despite the
anticipated use of CPME in the next step of the synthesis because of limited commercial
supplies and the relatively high cost of CPME. Methylation was complete (<0.3 of 5) after
6
7
8
o
stirring at ꢀ5 C for 0.5ꢀ1 h. The reaction was quenched by the addition of water over a
period of 45 min to control the exotherm and evolution of methane.17 After an extractive
workup, the anisole was removed18 azeotropically using water added in several portions.
After a solvent swap, the product was crystallized from cumene/heptane to give 8 as a
white crystalline solid in 81ꢀ85% yield on pilotꢀplant scale, with 100% purity by HPLC
assay and containing <5 ppm residual Ti and <1 ppm residual Li.
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
Table 1: Methylation of 5 using in situ-generated CH3TiCl3 from TiCl4 and CH3Li
a
Entry CH3TiCl3 Solventb
Temp
ºC
Time Product Unreacted Aldol
equiv
h
8
5
Diold
A%
0
A%c
98.2
98.6
88.0
A%
0.6
0.3
4.0
1
2
3
1.3
1.5
1.5
CPME
CPME
2ꢀ
ꢀ3 to 9
0 to ꢀ3
ꢀ5 to 8
6
3
16
0
5.1
MeTHF
MTBE
Bu2O
DME
THF
iꢀPr2O
iꢀPr2O
anisole ꢀ7 to ꢀ10
4
5
6
7
8
2.0
1.5
1.5
1.5
1.5
2.0
1.1
1.2
1.5
1.5
ꢀ5 to 8
ꢀ1 to 7
ꢀ7 to 9
6 to ꢀ7
1 to ꢀ4
0 to ꢀ4
9
7
7
7
5
2
4
2
1
1
76.2
21.5
73.2
46.1
97.7
99.5
94.2
99.3
99.7
97.8e
21.1
77.7
16.9
32.8
1.0
0.2
5.8
0.7
0.3
0
0
0
2.8
13.3
0
0
0
0
0
9
10
11
12
13
anisole
anisole
ꢀ3 to ꢀ5
ꢀ4 to ꢀ5
20ꢀ21
anisole
0
o
b
c
a CH3TiCl3 formed at ꢀ5 to ꢀ10 C from equimolar quantities of TiCl4 and CH3Li; Typically 10 V; Area
% by HPLC analysis. d Diol from methylation of aldol byꢀproduct 9. e Plus 1.3% of cuminyl impurities
12/13 and 0.6% of other impurities.
17 Alternatively, excess reagent can be quenched using acetone without methane evolution.
18 The carbinol 8 is highly soluble in anisole and in ether solvents in general. Attempts to isolate 8 from
anisole in high recovery by direct crystallization were not successful.
9 of 15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 15
1
2
3
Intramolecular Chichibabin cyclization
4
5
6
7
8
9
It had been known from prior experiences with synthesizing diazaindoles for related
chemical series that HMDS bases and LDA would be preferred bases to prepare 1 via an
intramolecular Chichibabin protocol.19 At least three equivalents of base were required for
the overall transformation, namely, to produce the alkoxide of 8, generate the benzylic
anion of 4 for reaction at the nitrile, and then to regenerate the anion at the benzylic
position to effect the cyclization. 20 KHMDS generally gave the best conversions.
NaHMDS had lower conversions even at longer reaction times. LiHMDS and LDA both
gave sluggish reactions and produced >6% of the uncyclized enamine 16.
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 highest conversions were obtained when the reaction was performed in ether
solvents. The initial two batches in a 1ꢀkg campaign of 1 employed solid KHMDS (4
equiv) in MTBE and 1.5 equiv of 2ꢀpropylpyrazine at 20ꢀ25 °C to give isolated yields of
55% and 72% of crude 1 after 24 h and 120 h, respectively. The timeꢀdependent increase
in yield was attributed to a partially reversible formation of benzamidine 14, an adduct
derived from 8 and KHMDS, likely as a sterically hindered monoꢀ or bisꢀtrimethylsilylated
form preventing trimerization to the sꢀtriazine. A maximum formation of 30% of 14 was
observed after 24 h that decreased to 18% after five days with a corresponding increase in
the formation of 1. The reaction conversion and isolation was facilitated by the
insolubility of the product. Thus, 1 was obtained in 72% yield in >97.5% purity by HPLC
analysis after simply diluting the reaction suspension with water and filtering.
19
Astolfi, D.; Ayers, T. A.; Chandramouli, S. V.; Hillegass, A.; Kubiak, G. G.; Lee, G. E.; Lythgoe, D. J.;
Powers, M. R.; Subotkowski, W.; Vanasse, B.; Weiberth, F. J.; Yu, Y. PCT Int. Appl. WO 2008/014249 A2,
2008.
20 For insights on the reaction mechanism, see ref. 2 and (a) Davis, M. L.; Wakefield, B. J.; Wardell, J. A.
Tetrahedron 1992, 48, 939. (b) McGill, C. K.; Rappa, A. Adv. Heterocycl. Chem. 1988, 44, 1.
10 of 15
ACS Paragon Plus Environment

Page 11 of 15
Organic Process Research & Development
1
2
3
Fig. 1. Side products.
4
5
OH
6
7
8
NH2
NH
N
N
9
N
NH2
N
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
HO
14
15
16
Because the solid KHMDS that was used in initial campaigns was found to have
inconsistent quality between lots and limited availability, the use of commercially
available KHMDS as a 19 wt% solution in THF was evaluated. Unfortunately,
unacceptable levels (4ꢀ8%) of the isopropyl side product 15 were formed in reactions
performed in THF, MeTHF, or alternative solvents that became enriched with THF that
was contributed from the KHMDS/THF reagent. The extent of the formation of 15 from 1
was a function of the solubility of 1 in the solvent system together with reaction
temperature and reaction time. A mitigating strategy that was successful in minimizing the
formation of 15 was employing CPME as a coꢀsolvent and reducing the amount of THF in
commercial KHMDS/THF to acceptable performance levels (<5 wt%) by coꢀdistillation
prior to reaction. Thus, CPME emerged as the preferred solvent for the cyclization, and
ultimately, a sourcing partner was developed that was able to supply bulk quantities of
KHMDS as a 32 wt% solution in CPME.
Further efforts to refine the Chichibabin cyclization centered on optimizing reaction
temperature, equivalents of 4 and equivalents of base. Yields improved to 87% by
increasing the reaction temperature from 20 ºC to 75 ºC likely because of a shift in the
amidine ↔ nitrile equilibrium favoring nitrile as temperature increased followed by
irreversible conversion of reformed nitrile to 1. At higher temperatures, an added benefit
was a decrease in reaction time from 24 h to 5 h. Employing 1.25 equiv of 4 was slightly
beneficial compared to using 1.1 equiv (87% vs. 83%). In a series of experiments using
o
1.25 equiv of 2ꢀpropylpyrazine and comparing reaction performance at 70 C using 3.1,
3.25, 3.5 and 4.0 equiv of KHMDS in CPME, 1 was obtained in 77, 81, 84 and 87%
isolated yield, respectively, after 5 h at 70 ºC. Temperatures much above 75 ºC were
avoided in order to minimize formation of the isopropyl side product 15. These parameters
11 of 15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 15
1
2
3
4
5
were successfully scaled in a pilot plant using 162.5 moles of 8, 1.25 equiv of 4, and 4.0
o
equiv of KHMDS in CPME. The reaction was complete after 6 h at 70ꢀ75 C. After
cooling, quenching with water, filtering and drying, 1 was obtained as lightꢀyellow solid in
6
7
8
81% yield.
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
In summary, the chemical development of a 2,3ꢀdisubstitutedꢀ4,7ꢀdiazaindole has been
described. The tertiary carbinol 8 was prepared employing in situꢀgenerated CH3TiCl3 as a
preferred reagent compared to CH3MgX for methyl addition to an enolizable ketone. The
4,7ꢀdiazaindole ring system was efficiently assembled via an intramolecular Chichibabin
transformation. The optimized processes were performed on pilotꢀplant scale to provide
72 kg of 1 in two campaigns in 60ꢀ68% overall yield.
Scheme 7. Final Synthesis
N
N
N
4
N
CH3TiCl3
N
N
O
HO
KHMDS
CPME
72 oC
anisole
-5 oC
80-86%
N
OH
H
75-82%
5
8
1
12 of 15
ACS Paragon Plus Environment

Page 13 of 15
Organic Process Research & Development
1
2
3
4
5
Experimental Section
Preparation of 4-(1-hydroxy-1-methylethyl)benzonitrile (8).
6
7
8
9
o
A reactor containing anisole (85 kg) was cooled to ꢀ10 to ꢀ5 C. TiCl4 (17 kg, 89.6 mol,
1.5 equiv) was charged over 1 h while maintaining ꢀ10 to ꢀ5 oC. After stirring at ꢀ10 to ꢀ5
oC for 15 min, an orange suspension was obtained. A solution of CH3Li in cumene/2ꢀ
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
MeTHF (61 kg, 3.2 wt%, 88.9 mol, 1.5 equiv) was charged over 100 min at ꢀ10 to ꢀ5 C.
o
The reddish solution was stirred at ꢀ10 to ꢀ5 C for 35 min. A solution of 5 (8.5 kg, 58.6
o
mol) in anisole (43 kg) was added over 80 min at ꢀ10 to ꢀ2 C. The reddish solution was
stirred at ꢀ5 to ꢀ2 oC for 60 min. HPLC analysis of an inꢀprocess sample indicated that the
reaction was complete (<0.5% 5). The reaction was quenched by the addition of water17,21
(4.3 kg) over 30 min at ꢀ10 to 0 oC. The reaction was partitioned by adding water (38 kg),
o
and was then warmed to 18 C. The organic phase was separated, and then washed with
water (2 x 43 kg), aqueous NaHCO3 (8% w/v, 2 x 110 kg) and water (2 x 43 kg). Anisole
was removed as an azeotrope with water (total 255 kg water added over five portions as the
o
reactor volume allowed) by distillation (45ꢀ50 Torr, <55 C internal temperature). The
resulting mixture, comprised of crude product as an organic phase (containing <0.5%
anisole by GC analysis) and a water phase, was diluted with cumene (22 kg), MTBE (31
kg) and aq NaCl (25 wt%, 20 kg), and then separated. The organic phase was then
o
partially concentrated (70ꢀ75 Torr, <60 C internal temperature) to remove MTBE until a
volume of about 30 L was achieved. The organic layer was filtered (0.8 µ cartridge) to
o
remove extraneous solids. The filtrate was heated to 65ꢀ70 C and heptane (5.8 kg) was
o
added. The clear solution was cooled to 20 C and seeded with pure 8 (8.5 g).
Crystallization commenced within 5 min exothermically (temperature rose to 23 °C in 20
o
min). The suspension was stirred at 20ꢀ22 C for 2 h, and then heptane (29 kg) was added
o
over a 60ꢀmin period. The suspension was stirred at 20ꢀ22 C for 1 h, cooled and aged at
o
0ꢀ5 C for 3 h, and was then filtered. The cake was washed with cumene/heptane (8.5
kg:0.5 kg) and heptane (8.5 kg) and then dried under reduced pressure at 35 °C to afford
21
Gas evolution was vigorous during the initial part of the quench.
13 of 15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 15
1
2
3
1
8.0 kg (84% yield, 100% purity22) of 8 as a white crystalline solid. H NMR (300 MHz,
4
5
CDCl3): 1.58 (s, 6H), 1.85 (s, 1H), 7.45ꢀ7.55 (m, 4H).
6
7
Preparation of 2-[4-(7-ethyl-5H-pyrrolo[2,3b]pyrazin-6-yl)phenyl]propan-2-ol (1)
8
9
A reactor was charged with 8 (26.2 kg, 162.5 mol), 2ꢀpropylpyrazine (25.0 kg, 204.6 mol,
1.25 equiv) and CPME (180 kg). Potassium bis(trimethylsilyl)amide (410 kg of 32 wt%
solution in CPME, 657 mol, 4 equiv) was added over 70 min while allowing the
temperature to rise to about 35 oC. The batch was heated and held at 70ꢀ76 oC for 6 h. The
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
suspension was cooled to 22 C over 1 h. Water (81 kg) was added in 45 min while
maintaining <45 oC. The thick suspension was stirred for 35 min, cooled and aged at 5 oC
for 1.5 h, and then was transferred to a filter followed by CPME/water rinse (41:7 kg).
The cake was washed with water/MeOH (580 kg/200 kg) followed by water (to pH < 12),
o
and was then dried at 70 C to afford 1 as a lightꢀyellow solid (37.3 kg, 132.5 mol, 81.6%
1
yield, 99.6% purity23). H NMR (300 MHz, DMSOꢀd6): 1.29 (t, J = 7.4 Hz, 3H), 1.48 (s,
6H), 2.92 (q, J = 7.5 Hz, 2H), 5.12 (s, 1H), 7.64 (s, 4H), 8.20 (d, J = 2.6 Hz, 1H), 8.35 (d, J
= 2.6 Hz, 1H). 12.03 (br s, 1H)
Acknowledgements
We gratefully acknowledge Andrew Bridge, David Lythgoe and Michel Mulhauser for
their encouragement and helpful discussions, and Geoffrey Darnbrough, Anthony
Gardetto, Matthew Powers, Laura Turci, Benoit Vanasse and other colleagues within our
organization who provided process safety, analytical, sourcing and pilot plant operations
support.
22
o
Zorbax Eclipse XDBꢀC8, 4.6 x 150 mm, 5 ꢁ, 240 nm, 25 C, water/ACN/TFA, 1 mL/min, gradient
program: 70/30/0.1 for 10 min, then linear ramp over 5 min to 15/85/0.1, then held for 5 min. Relative
retention times: 8, 1.00; 5, 1.33
.
23
o
Zorbax Eclipse XDBꢀC8, 4.6 x 150 mm, 5 ꢁ, 250 nm, 35 C, 40:60:0.1 water/ACN/TFA isocratic, 1
mL/min for 20 min. Relative retention times: 1, 1.00; 8, 1.23
.
14 of 15
ACS Paragon Plus Environment

Page 15 of 15
Organic Process Research & Development
1
2
3
4
5
6
7
Development of a Synthesis of a 2,3-
Disubstituted-4,7-Diazaindole
Including Large-Scale Application of
CH3Li/TiCl4-Mediated Methylation
of an Enolizable Ketone
N
Cl
N
n-PrMgCl
cat.
N
N
7
4
N
N
base
N
OH
H
N
N
8
9
Franz J. Weiberth,* Harpal S. Gill, George
E. Lee, Duc P. Ngo, Frederick L. Shrimp II,
Xuemin Chen, Geoffrey D’Netto, Bryan R.
Jackson, Ying Jiang, Narendra Kumar,
Frederick Roberts and Evgeny Zlotnikov
1
CH3MgBr
O
HO
or
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
"CH3TiCl3"
5
8
* Author to whom correspondences should be sent via email: franz.weiberth@sanofi.com
15 of 15
ACS Paragon Plus Environment