Development of a Practical Synthesis of ERK Inhibitor GDC-0994

Article abstract of DOI:10.1021/acs.oprd.7b00006

The process development of a synthetic route to manufacture ERK inhibitor GDC-0994 on multikilogram scale is reported herein. The API was prepared as the corresponding benzenesulfonate salt in 7 steps and 41% overall yield. The synthetic route features a biocatalytic asymmetric ketone reduction, a regioselective pyridone S_N2 reaction, and a safe and scalable tungstate-catalyzed sulfide oxidation. The end-game process involves a telescoped S_NAr/desilylation/benzenesulfonate salt formation sequence. Finally, the development of the API crystallization allowed purging of process-related impurities, obtaining >99.5A% HPLC and >99% ee of the target molecule.

Full text of DOI:10.1021/acs.oprd.7b00006

Subscriber access provided by University of Newcastle, Australia
Full Paper
Development of a Practical Synthesis of ERK Inhibitor GDC-0994
Xin Linghu, Nicholas Wong, Hans Iding, Vera Jost, Haiming
Zhang, Stefan G. Koenig, C. Gregory Sowell, and Francis Gosselin
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.7b00006 • Publication Date (Web): 23 Feb 2017
Downloaded from http://pubs.acs.org on February 23, 2017
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 40
Organic Process Research & Development
1
2
3
4
5
6
Development of a Practical Synthesis of ERK
Inhibitor GDC-0994
7
8
9
Xin Linghu,*,† Nicholas Wong, † Hans Iding, ‡ Vera Jost, ‡ Haiming Zhang, †
Stefan G. Koenig, † C. Gregory Sowell,† and Francis Gosselin†
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
†Small Molecule Process Chemistry, Genentech, Inc., 1 DNA Way, South San Francisco,
California 94080, United States
‡Process Research, F. HoffmannꢀLa Roche AG, Grenzacherstrasse 124, CHꢀ4070 Basel,
Switzerland
Corresponding Author
*Eꢀmail: linghux@gene.com.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 40
1
2
3
4
For Table of Contents Only
5
Cl
6
7
8
9
N
N
O
O
Cl
F
F
O
Cl
F
HN
N
S
N
O
N
HO
+
O
N
N
BsOH
GDC-0994
41% yield in 7 steps
>99.5%purity, 99.5% ee
N
N
OH
OTBS
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
NH
MeS
N
ACS Paragon Plus Environment

Page 3 of 40
Organic Process Research & Development
1
2
3
4
5
ABSTRACT
The process development of a synthetic route to manufacture ERK inhibitor GDCꢀ0994
6
7
8
9
on multiꢀkilogram scale is reported herein. The API was prepared as the corresponding
benzenesulfonate salt in 7 steps and 41% overall yield. The synthetic route features a
biocatalytic asymmetric ketone reduction, a regioselective pyridone SN2 reaction and a
safe and scalable tungstateꢀcatalyzed sulfide oxidation. The endꢀgame process involves a
telescoped SNAr / desilylation / benzenesulfonate salt formation sequence. Finally,
development of the API crystallization allowed purging of processꢀrelated impurities,
obtaining >99.5A% HPLC and >99% ee of the target molecule.
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: biocatalytic asymmetric ketone reduction; regioselective SN2 of pyridone;
tungstateꢀcatalyzed oxidation; SNAr displacement of sulfone; oneꢀpot desilylation / salt
formation.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 40
1
2
3
4
1. INTRODUCTION
5
6
7
8
The extracellularꢀsignalꢀregulated kinases (ERK1 and ERK2) represent an essential
node within the RasꢀRafꢀMEKꢀERK signaling cascade that controls a number of
fundamental cellular processes including cell survival, proliferation, motility, and
differentiation.1 Once activated, ERK1/2 kinases phosphorylate several downstream
targets involved in cytoskeletal changes and transcriptional activation.2ꢀ5 Furthermore,
mutations of upstream genes (Ras and BRAF) that involve a pathway reactivation
through ERK1/2 have been identified in many human tumors.6ꢀ8 These observations have
promoted the development of several small molecule inhibitors targeting ERK1/2
kinases.9ꢀ15 GDCꢀ0994 (Figure 1) was discovered as a potent, selective, and efficacious
inhibitor of ERK1/2 kinases.16,17 We describe herein an efficient multiꢀkilogram scale
asymmetric synthesis of GDCꢀ0994 to support the preclinical and clinical studies.
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. Structure of ERK inhibitor GDC-0994
N
O
Cl
F
HN
N
N
N
N
BsOH
GDC-0994
OH
2. DISCUSSION AND RESULTS
Discovery Synthesis. The initial discovery synthesis of GDCꢀ0994 freeꢀbase is
outlined in Scheme 1. This route comprises 7 synthetic steps and the target molecule was
produced in 12% overall yield. Chiral diol 2 was generated via an osmiumꢀcatalyzed
asymmetric Sharpless dihydroxylation of styrene 1. After selective monoꢀprotection as a
TBSꢀether and mesylation of diol 2, the SN2 reaction between pyridone 518 and secondary
methanesulfonate 4 was able to produce sulfide 6. Sulfone 7 was obtained via an
oxidation of 6 with m-CPBA. Finally, GDCꢀ0994 freeꢀbase 10 was synthesized through a
ACS Paragon Plus Environment

Page 5 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
sequence of SNAr displacement of sulfone 7 with commercially available aminopyrazole
8, followed by desilylation. While this route enabled the production of hundredꢀgram
quantities of GDCꢀ0994 freeꢀbase for preclinical studies, the chemistry suffered from low
yields and modest selectivity in several steps. In addition, tedious purification procedures
such as distillation and flash chromatography were required due to fairly complex
reaction profiles. Issues revolving around process efficiency, reaction safety and
robustness deterred the chemistry from being used on multiꢀkilogram 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 1. Discovery synthetic route to GDC-0994 free-base
Efficient Asymmetric Synthesis of Chiral Diol. In the original route, chiral diol 2
was synthesized from styrene 1 through a Sharpless asymmetric dihydroxylation. The
chiral diol was obtained in 94% ee, a result acceptable for early stage research, but not for
our clinical quality requirements. While chiral purity could not be further upgraded in the
downstream chemistry and isolations, a selective crystallization of racemic GDCꢀ0994
freeꢀbase allowed for the desired (S)ꢀenantiomer to reach 98.9% ee in the mother liquor
(Scheme 2). Although this process could upgrade chiral purity to an acceptable level, the
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 40
1
2
3
4
5
6
7
8
9
inherent mass loss of the desired product rendered it less efficient. In addition to the
optical purity issue, tedious purifications were involved in this route, which were
considered nonꢀoptimal and preferably removed from the process.
Scheme 2. Chiral purity upgrade of GDC-0994 free-base via a selective crystallization
of API racemate
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
Ketoreductase (KRED)ꢀmediated biotransformations have proven exceptionally
useful and often a preferred method of chiral alcohol synthesis.19 To circumvent the
limitations of the dihydroxylation chemistry, we envisioned that a biocatalytic reduction
of αꢀhydroxyketone 1120 would provide efficient access to the key diol intermediate 2
(Scheme 3). 21
Scheme 3. Chiral diol synthesis via enzymatic ketone reduction
We initiated a focused screen for the asymmetric reduction of αꢀhydroxyketone 11
based on our prior experience with biocatalysis on similar substrates.22 Active (R)ꢀ
enantioselective KREDs were screened, applying two different cofactorꢀrecycling
systems. One utilized 2ꢀpropanol (Table 1) while the other used Dꢀglucose (Table 2) as
the terminal reductant. Both reagents were suitable for the use of the cofactors, NADH
ACS Paragon Plus Environment

Page 7 of 40
Organic Process Research & Development
1
2
3
4
5
6
and NADPH. In the latter case, an auxiliary enzyme, glucose dehydrogenase (GDH) was
required for cofactor regeneration.
7
8
9
Table 1: Enzymatic ketone reduction with 2-propanol as reductanta
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
F
Cl
F
ketoreductase
HO
HO
O
2
11
HO
NAD(P)H
NAD(P)
acetone
2-propanol
ketoreductase
Entry
Enzyme
% Conversion
% ee
1
2
3
4
5
6
P2ꢀG03
P3ꢀB03
P3ꢀG09
P3ꢀH12
NADHꢀ112
A131
99.4
11
99.2 (S)
94.6 (R)
84.0 (R)
97.9 (R)
92.5 (R)
56.9 (R)
5.8
30.4
10.3
1.7
a Screening conditions: 11 (0.053 ꢀmol), 2 mM MgCl2, 0.1 M potassium phosphate buffer, pH 7.2;
NAD or NADP s/c = 100; enzyme s/e = 10; 5% (v/v) 2ꢀpropanol; 5% (v/v) acetonitrile ; rt.
While the singleꢀenzyme system resulted in low conversions or formation of the
undesired (S)ꢀenantiomer (Table 1), KRED NADHꢀ112 with a Dꢀglucose / GDH
recycling system was identified as a highlyꢀefficient catalyst, producing the desired diol 2
with >99% ee and 100% conversion (Table 2, entry 5). The screening result was first
confirmed on gram scale at a 5% (w/v) concentration where the reaction was completed
in 3 days at ambient temperature, producing 2 in 98% isolated yield and >99.9% ee.
Further optimization of the reaction conditions allowed the substrate concentration to be
increased to 10% (w/v). The reaction temperature was increased to 30 °C in order to
shorten the conversion time to less than 24 h.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 40
1
2
3
4
Table 2: Enzymatic ketone reduction with D-glucose as a reductanta
5
6
7
8
Cl
F
Cl
F
ketoreductase
HO
HO
O
2
11
HO
NAD(P)H
NAD(P)
9
glucose
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
gluconic acid
glucose dehydrogenase
Entry
Enzyme
% Conversion
% ee
1
2
3
4
5
6
P2ꢀG03
P3ꢀB03
P3ꢀG09
P3ꢀH12
NADHꢀ112
A131
99.8
2.7
0.9
12.4
98
100 (S)
71.5 (R)
40.8 (R)
95.4 (R)
99.5 (R)
85.4 (R)
5.7
a Screening condition: 11 (0.1 ꢀmol), 2 mM MgCl2, 0.1 M potassium phosphate buffer, pH 7.2;
0.25 M D-glucose; GDHꢀ105 s/e = 100 and NAD or NADP s/c = 100; enzyme s/e = 5; rt.
Methanesulfonate Formation. A twoꢀstep, oneꢀpot procedure was developed to
synthesize methanesulfonate 4 (Scheme 4). Selective protection of the primary alcohol
with a TBS group proceeded smoothly with minimal formation of the overꢀprotected byꢀ
product. We initially used Et3N with a catalytic amount of DMAP as base for both steps.
The reaction was heterogeneous in dichloromethane at a 7% (w/v) concentration and
went to completion in over 12 h at 20 °C. Substitution of Et3N with Hünig’s base allowed
reaction completion in 2 h at the same temperature and remained homogenous at a 20%
(w/v) concentration. Methanesulfonyl chloride was then introduced to the reaction
mixture to yield 4 in >98A% HPLC purity. Due to the instability of 4, after workꢀup, the
crude solution was taken to the next step without further purification.
ACS Paragon Plus Environment

Page 9 of 40
Organic Process Research & Development
1
2
3
4
Scheme 4. One-pot synthesis of methanesulfonate 4
5
6
7
8
Cl
F
Cl
F
1. DIPEA/DMAP,
TBSCl in DCM
MsO
HO
HO
2. MsCl
TBSO
4
2
87% over 2 steps
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
SN2 Displacement. With access to the methanesulfonate 4 secured, we turned our
attention to the SN2 reaction with pyridone 5 (Scheme 5).23 In the original process, the
reaction produced Nꢀalkylation adduct 6 in modest 50ꢀ60% yields along with >20% of
the undesired Oꢀalkylation byꢀproduct 12.
Scheme 5. SN2 reaction of methanesulfonate 4 and pyridone 5
A series of reaction conditions were evaluated to improve the regioselectivity in the
alkylation reaction. Weak organic (TMG, DABCO, 2,6ꢀlutidine and DMAP) and
inorganic bases (alkali metal fluorides, carbonates and phosphates) resulted in <60%
conversions to 6 in 24 h. In a few cases (DBU and PPh3 / DIAD), the reaction went to
completion, but favored the undesired Oꢀalkylation byꢀproduct 12. While some strong
bases (metal hydride and alkali metal tertꢀbutoxide) afforded messy reaction profiles,
metal hexamethyldisilazide (MHMDS) variants in general led to a cleaner reaction and
favored formation of the desired Nꢀalkylation product 6. The counterꢀion effect on the
regioselectivity of the SN2 reactions was then investigated (Table 3). By comparison of
lithium, sodium and potassium, it appeared that as the size of counterꢀion increased, the
selectivity was improved. However, the cesium salt of pyridone 5 deviated from this
trend and gave a low selectivity.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 40
1
2
3
4
Table 3. Counter-ion effect on regioselectivitya,b
5
N vs O ratio
6:12b
6
7
8
9
Entry
MHMDS
% Conversionb
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
2
3
4
LiHMDS
NaHMDS
KHMDS
90
>95
>95
92
71:29
79:21
84:16
Cs salt of 5c
73:27
a Reaction conditions: 1.0 equiv of 4 (5 mmol), 1.1 equiv of 5 (5.5 mmol), 1.1 equiv of MHMDS
in THF, 80 °C for 8 h; b Conversion and selectivity were monitored by HPLC analysis. c Cs salt
of 5 was prepared by mixing a 1:1 ratio of CsOH solution and 5 followed by precipitation with
CH3CN.
Solvent also played a critical role in the regioselectivity of the SN2 reaction (Table
4). Although polar solvents allowed for the reaction to go to completion in a few hours,
there was no differentiation between two alkylation pathways, resulting in 1:1 mixtures
of both regioisomers (entries 1ꢀ3). On the other hand, nonꢀpolar solvents in general
enable a reaction with reasonable selectivity (entries 4ꢀ6). However, due to the low
solubility of 4, thick slurries were obtained and made the agitation challenging. The
addition of 5ꢀ10% NMP as a coꢀsolvent helped to alleviate the mixing issue. We
observed that the reaction stalled at ~90% conversion in THF / NMP or toluene / NMP
and required extended reaction times to achieve full conversion. The reaction in 1,4ꢀ
dioxane with 5% NMP gave a very thin slurry. As a result, full conversion was obtained
in 18 h with an acceptable level of N vs O selectivity (entry 7). Fortunately, the Oꢀ
alkylated regioisomer 12 was purged below 0.5A% HPLC after crystallization of 6 from
nꢀheptane. Under all those reaction conditions, a complete inversion of stereochemical
ACS Paragon Plus Environment

Page 11 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
configuration of the benzylic chiral center was observed during the displacement reaction
and no erosion of enantiomeric excess could be detected based on HPLC analysis of the
crude reaction mixture.
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 4. Solvent effect on the conversion and regioselectivitya
N vs O
Co-
Temperature
(°C)
%
Entry
Solvent
ratio
6:12b
50:50
49:51
58:42
84:16
88:12
89:11
87:13
solvent
Conversionb
1
2
3
4
5
6
NMP
DMF
ꢀ
100
100
70
100
100
100
89
ꢀ
MeCN
toluene
THF
ꢀ
10% NMP
10% NMP
ꢀ
100
70
95
1,4ꢀdioxane
70
93
7
1,4ꢀdioxane 5% NMP
70
100
a
Reaction conditions: 1.0 equiv of 4 (5 mmol), 1.1 equiv of 5 (5 mmol), 1.1 equiv of KHMDS in
THF; at the indicated temperature for 18 h; b Conversion and selectivity were monitored by HPLC
analysis.
Sulfone Synthesis via Oxidation. Conversion of sulfide to the sulfone was initially
accomplished with 3ꢀchloroperbenzoic acid (m-CPBA) in dichloromethane (Scheme 6).
Sulfone 7 was isolated in 80ꢀ85% yield as a pure crystalline solid. These reaction
conditions were then scaled up to produce the first few batches of 7. Although the
reaction yields were consistently high, pyrimidine Nꢀoxide impurity 13 (~1A% HPLC)
was observed and led to the corresponding Nꢀoxide impurity in the final API (~0.5A%
HPLC).24 Unfortunately, there was no purge of API Nꢀoxide in the final salt formation
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 40
1
2
3
4
5
6
7
8
9
and recrystallization steps. The rejection of 13 after the oxidation step was attempted via
a crystallization of 7 in THF / n-heptane but the purge efficiency proved to be quite low
(less than oneꢀfold). As a consequence, four recrystallizations were required to reduce the
level of 13 from 0.8% to below 0.15%. In addition to the impurity issue, the m-CPBA
reaction was deemed to be an environmentally unfavorable process due to a low atom
economy with formation of a benzoic acid byꢀproduct and use of undesirable
dichloromethane as reaction solvent.
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. Sulfone synthesis via an m-CPBA oxidation
Table 5. Solvent effect on sulfide oxidationa
N
30% H2O2,
Na2WO4 (1 mol%)
solvent
N
N
O
Cl
F
O
Cl
F
O
Cl
F
MeS
N
S
N
+
S
N
O
N
O
N
O
N
6
7
14
OTBS
OTBS
OTBS
A%
A%
A%
Entry
Solvent
Time (h)
HPLC 6 HPLC 14 HPLC 7
1
2
3
4
5
6
7
EtOH
7
3
15
1
18
1
67
98
EtOH / H2O (9:1)
NMP / H2O (9:1)
DMF / H2O (9:1)
THF / H2O (9:1)
22
22
22
6
28
10
6
1
71
5
85
3
91
MeCN / H2O (9:1)
nꢀPrOH / H2O (50:1)
1
1
98
2
1
0
>99
ACS Paragon Plus Environment

Page 13 of 40
Organic Process Research & Development
1
2
3
4
a Reaction conditions: 1.0 equiv of 6 (5 mmol), 2.2 equiv of 30 wt% H2O2, 1 mol% of Na2WO4, in 10
5
6
mL/g of solvent
7
8
9
As an alterative to m-CPBA, aqueous hydrogen peroxide has been shown in the
literature to be one of cheap and green oxidants.25 Sulfone synthesis via a molybdateꢀ or
tungstateꢀcatalyzed sulfide oxidation with aqueous hydrogen peroxide has been well
documented.26ꢀ28 We then quickly evaluated the efficiency of Na2WO4ꢀcatalyzed
oxidation of sulfide 6 with 30% aqueous H2O2 in a variety of solvents (Table 5). It was
found that EtOH (entry 1) gave an incomplete conversion after 7 h. Reactants were not
completely soluble in this solvent system, resulting in a heterogeneous reaction. With the
addition of 10% (v/v) of water, all reagents were soluble and full conversion was
observed after 3 h (entry 2). Polar aprotic solvents NMP, and DMF (entries 3 and 4) were
found to have greater solubility but significantly decreased the rate of reaction, resulting
in incomplete reaction after extended periods of time. Aqueous THF or acetonitrile
(entries 5 and 6) were found to have excellent solubility of reagents and products,
resulting in a 90ꢀ98% conversion. nꢀPrOH with 2% (v/v) of water (entry 7) was identified
to be optimal, allowing for >99% conversion in 2 h.
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. Observed impurities in sulfide oxidation with H2O2
N
N
O
Cl
F
O
Cl
F
HO
N
S
N
O
N
O
N
16
15
OTBS
OH
Due to the acidic nature of H2O2,29 desilylated sulfone byꢀproduct 15 was observed
within 2 h reaction time at up to 3.2A% HPLC (Table 6, entry 1). A stress test indicated
that nearly full desilylation was obtained after 18 h under the reaction conditions (entry 2).
A base screen was performed to suppress formation of 15 (Figure 2). While other bases
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 40
1
2
3
4
5
6
7
8
9
had either a low efficiency to minimize 15 or formed more hydrolysis impurity 16
(entries 3ꢀ5), we found that 2,6ꢀlutidine (entry 6) was efficient to minimize desilylation
without significantly increasing the level of 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
Table 6. Base additive to minimize desilylation and hydrolysis impuritiesa
Entry
Base
ꢀ
mol% Time (h) A% HPLC 15
A% HPLC 16
1
2
3
4
5
6
ꢀ
2
18
2
3.2
95
0.50
0.50
0.50
3.0
ꢀ
ꢀ
iPr2EtN
K2CO3
KOAc
2,6ꢀLutidine
10
10
10
10
8
2
1.4
2.5
0.28
2
0.90
0.53
2
a
Reaction conditions: 1.0 equiv of 6 (5 mmol), 2.2 equiv of 30 wt% H2O2, 1 mol% of Na2WO4, in 10
mL/g of nꢀPrOH / H2O (50:1)
It is well known that there are potential process hazards associated with a tungstateꢀ
catalyzed oxidation with hydrogen peroxide. This transformation is susceptible to
exothermic decomposition of H2O2 leading to runaway reactions and potential evolution
of oxygen. The reaction safety aspects for a tungstateꢀH2O2 oxidation system have been
described in some recent publications.30,31 With optimal reaction conditions established,
we then focused on the reaction safety and scalability, addressing potential issues with
heat and peroxide accumulation in the oxidation reaction.
Two reaction conditions were examined. When H2O2 was charged to the reaction
mixture at 40 °C over 2 h, 45% conversion of sulfide 6 was observed at the end of the
addition with formation of a mixture of sulfone 7 and sulfoxide intermediate 14. This
result indicated a high level of peroxide accumulation after charging as the reaction went
ACS Paragon Plus Environment

Page 15 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
to completion within additional 4ꢀ5 hours. Reaction calorimetry was carried out to
measure the enthalpy of the reaction in a reaction calorimeter (RC1), revealing a
significant heat accumulation after charging (Table 7). This translated to a high adiabatic
temperature rise of 12.7 °C. In contrast, when H2O2 was added to a reaction mixture at
80 °C with the same addition time, >95% conversion of sulfide to sulfone was observed.
As a result of the faster reaction rate, minimal H2O2 was accumulated in the reaction and
the postꢀaddition enthalpy was significantly reduced.
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 7. Temperature effect on heat and H2O2 accumulationa
Reaction and
thermal parameter
Conversion of sulfide
at the end of addition
Overall enthalpy of
reaction
Reaction
Reaction
temperature 40 °C
temperature 80 °C
45%
>95%
ꢀ444.1 kJ/mol
ꢀ282 kJ/mol
ꢀ162.1 kJ/mol
12.7 °C
ꢀ503.5 kJ/mol
ꢀ475.8 kJ/mol
ꢀ27.7 kJ/mol
3.1 °C
Enthalpy of reaction
during H2O2 addition
Enthalpy of reaction
postꢀaddition
Adiabatic temperature
rise postꢀaddition
With a safe operation procedure in place, the oxidation was successfully
demonstrated and upon reaction completion, residual H2O2 was quenched with an
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 40
1
2
3
4
5
6
7
8
9
aqueous sodium ascorbate solution at 80 °C. The crystalline sulfone product was directly
precipitated from the reaction mixture with the addition of water as an antiꢀsolvent. The
crystallization was found to be effective in purging impurities 15 (3×) and 16 (>10×). The
product was obtained in 87% isolated yield with Nꢀoxide impurity 13 controlled below
the detection limit (LOD <0.05A% HPLC).
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
Sulfone Displacement with Aminopyrazole (SNAr Step). SNAr displacement
(Scheme 7) was initially carried out in DMF and employed Cs2CO3 as a base. The
reaction suffered from a modest conversion at ambient temperatures and a complex
reaction profile at 80 °C. The average yield of the reaction was approximately 60%. The
modest isolated yields were due to the formation of significant quantities of byꢀproduct
17 (Figure 3) via hydrolysis of the sulfone functionality (Table 8, entry 1). To improve
the reaction yield, the reactivity and selectivity issues would have to be resolved. We
speculated that the nucleophilicity of aminopyrazole 8 could be increased by a direct
deprotonation of 8 with a strong base. On the other hand, the formation of a high level of
17 could be attributed to a high concentration of metal hydroxide in the reaction solution,
leading to a competitive displacement of the sulfone. In contrast to the high solubility of
CsOH in DMF, LiOH has a much lower solubility in THF. We surmised that LiHMDS in
THF would be a good combination to not only increase the reactivity of 8 as an anion but
also reduce the hydrolysis side reaction.
Scheme 7. Sulfone displacement with aminopyrazole
ACS Paragon Plus Environment

Page 17 of 40
Organic Process Research & Development
1
2
3
4
Figure 3. Observed impurities in SNAr step
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
As shown in Table 8, when LiHMDS was employed as a base, the reaction rate was
dramatically increased. At ambient temperature, the reaction went to completion
immediately after the addition of base (entry 2). By comparison, <90% conversion was
obtained after 24 h using the original procedure (entry 1). The change in conditions also
decreased the level of hydrolysis byꢀproduct 17 from 15ꢀ20A% HPLC to 3A% HPLC.
Formation of 17 was further reduced to below 1A% HPLC after the reaction temperature
was decreased to –30 °C. In addition, the decrease in reaction temperature had no
noticeable impact on the reaction rate (entry 3). When 10 mol% of water was added to
the reaction solution, the same level of 17 was observed (entry 4) as that under the
anhydrous conditions. This observation indicated the same concentration of LiOH under
both anhydrous and wet conditions. This is presumably due to a low solubility of LiOH in
THF, leading to a precipitation of LiOH out of the saturated solution and a negligible
impact on the impurity formation.
Table 8: Base and solvent effect on SNAra,b
Temperature
%
Conversion
<90
A%
HPLC 17
15ꢀ20
Entry
Base
Solvent
Time
(°C)
1
2
Cs2CO3
DMF
THF
25
24 h
10
LiHMDS
25
100
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 40
1
2
3
4
min
5
6
7
8
9
3
4
LiHMDS
LiHMDS
THF
ꢀ30
ꢀ30
10 min
100
100
<1
<1
THF with
10 mol%
water
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
10 min
a Reaction conditions: 1.0 equiv of 7 (5 mmol), 1.2 equiv of 8 (6 mmol), 2.1 equiv of base at the indicated
temperature; b Reactions were monitored by HPLC analysis.
Although employing a strong base resulted in improved conversions and reaction
profiles, two new impurities were introduced. Due to a fairly acidic proton at the chiral
benzylic position, one impurity 18 (Figure 3) was generated through an elimination
pathway when using excess LiHMDS. The other impurity 19 (Figure 3) was derived from
18 through the subsequent addition of 8. The stoichiometry of base is critical of achieving
full conversion while minimizing the formation of 18 and 19 (Table 9). The NH proton of
the product 9 is the most acidic site in the reaction mixture and acts as an irreversible
base sink. As a result, only 50% conversion was observed when 1 equivalent of LiHMDS
was applied (entry 1). Since the NH proton of 8 is more acidic than the benzylic proton,
an excess of 8 is able to serve as a buffer for the overꢀcharged base. When slight excesses
of LiHMDS were employed and trapped by 8, no significant byꢀproducts were formed
(entry 2). In contrast, when large excesses of base were charged and cannot be
completely consumed by 8, a significant amount of 18 and 19 were observed (entry 3).
ACS Paragon Plus Environment

Page 19 of 40
Organic Process Research & Development
1
2
3
4
Table 9. Impact of LiHMDS equivalents on conversion and impurity formation
5
6
7
8
Li
N
N
N
H
H
N
H
O
H
Cl
F
O
Cl
F
TMS
TMS
N
N
N
S
N
O
N
O
N
+
THF, -40 o
C
N
N
9
N
9
OTBS
OTBS
7
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
1.0 equiv
1.2 equiv
Equivalent
A% HPLC
A% HPLC
Entry
% Conversion
of LiHMDS
18a
ꢀ
19a
ꢀ
1
2
3
1.0
2.1
2.3
~50
100
100
1.1
6.5
1.2
4.1
a Reaction conditions: 1.0 equiv of 7 (5 mmol), 1.2 equiv of 8 (6 mmol), indicated amount of LiHMDS at
–40 °C; b Reactions were monitored by HPLC analysis.
The level of both impurities 18 and 19 can be further reduced in the downstream
benzenesulfonate salt formation and API recrystallization steps. The measured purge
factors of 18 were 3× in the salt formation step and 2× in the final recrystallization step.
The rejections of 19 in those two steps were more efficient. To purge individual
impurities below the GMP specifications at the end, the level of impurity 18 needed to be
controlled below 0.5% in the SNAr step. This requirement led to further optimization of
bases, solvents and temperatures (Table 10). The base screen indicated that LiHMDS was
the best choice (entry 1). Other strong bases, KHMDS, NaHMDS or LiOtBu resulted in
either a higher level of both 18 and 19 or formation of other impurities (entries 2ꢀ4). Coꢀ
solvents had a significant impact on the level of impurities. While use of the polar coꢀ
solvent NMP was unsuccessful to reduce the level of both impurities (entry 5), nonꢀpolar
coꢀsolvent, MTBE, toluene and 2ꢀMeTHF improved the reaction purity profile (entries 6ꢀ
8). 0.60% of 18 and <0.1% of 19 was observed in the reaction when employing LiHMDS
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 40
1
2
3
4
5
6
in 2ꢀMeTHF (entry 9). The level of both impurities can be further reduced when the
reaction temperature was decreased to –25 °C (entry 10).
7
8
9
Table 10. Optimization of Conditions for SNAra
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%
HPLC
18b
A%
HPLC
19b
Entr
y
Solvent /
additive
Temp.
A%
Base
(°C) HPLC 9b
1
2
3
4
LiHMDSd
KHMDSd
NaHMDSd
LiOtBud
THF
THF
THF
THF
5
5
5
5
88
42
4.2
11
14
ꢀ
2.1
21
9.0
ꢀ
63
58c
THF / NMP
20:1 (v/v)
5
LiHMDSd
5
80
5.3
4.9
THF / MTBEg
1:1 (v/v)
6
7
8
LiHMDSd
LiHMDSe
LiHMDSd
5
5
5
91
92
89
1.1
2.8
2.7
0.12
1.1
THFh
THF / 2ꢀMeTHFg
(1:1 v/v)
1.5
9
LiHMDSf
LiHMDSf
THFi
THFi
5
94
96
0.60
0.31
<0.1
n/d
10
–25
a Reaction conditions: 1.0 equiv of 7 (5 mmol), 1.2 equiv of 8 (6 mmol), 2.1 equiv of base in the indicated
solvent, 10 mL/g of the final reaction volume after addition of a base solution; b Reactions were monitored
ACS Paragon Plus Environment

Page 21 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
by HPLC analysis; c The reaction contained 42A% HPLC of byꢀproduct with tꢀbutyloxide displacement of
sulfone functionality; d MHMDS or MOtBu in THF (1.0 M); e LiHMDS in toluene (1.0 M); f LiHMDS in 2ꢀ
MeTHF (0.5 M). g Final solvent ratio THF:MTBE or THF:2ꢀMeTHF = 1.7:1 (v/v). h Final solvent ratio
THF:toluene = 3:1 (v/v). i Final solvent ratio THF:2ꢀMeTHF = 1.5:1 (v/v).
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. Telescoped desilylation and benzenesulfonate salt formation
SO3H
N
N
N
N
O
Cl
F
HCl/MeOH
45 °C
O
Cl
F
O
Cl
F
HN
N
HN
N
HN
N
N
N
N
N
N
MEK/water
N
N
N
BsOH
GDC-0994
OH
9
10
OTBS
OH
71% over three
steps from 7
99.5% purity, >99.5% ee
individual impurity <0.25%
Telescoped Desilylation and Benzenesulfonate Salt Formation. No crystalline
forms were found for the freebase of GDCꢀ0994 10 following extensive screening. After
salt screens, a few crystalline salt forms were identified with strong acids (HCl, pꢀ
toluenesulfonic acid and benzenesulfonic acid). Among them, the benzenesulfonate salt
of GDCꢀ0994 possessed superior stability and very low hygroscopicity. As a result, it
was identified as an appropriate polymorph for further development.
The desilylation was initially achieved using HCl in MeOH at 45 °C (Scheme 8). The
reaction typically proceeded to 70ꢀ80% conversion within the first 3 h, but required more
than 10 h to go to completion. After the reaction was finished, an azeotropic distillation
was carried out to facilitate the removal of the corresponding byꢀproduct, tertꢀbutylꢀ
methoxyꢀdimethylsilane, before quenching. The content of MeOH in EtOAc was then
controlled at 15–20 wt% to ensure a good solubility of GDCꢀ0994•HCl salt in the organic
phase, while minimizing aqueous losses. Since the formation of benzenesulfonate salt
involved 1 equivalent of benzenesulfonic acid, there was a quality concern associated
with the introduction of the genotoxic impurity methyl benzenesulfonate as a byproduct
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 40
1
2
3
4
5
6
7
8
9
from reaction of the acid in methanol. MeOH levels were reduced to below 0.3A% GC
through concentration and solvent swap before the benzenesulfonate salt formation step.
GDCꢀ0994•benzenesulfonate has a low solubility in methyl ethyl ketone (MEK) (<5
mg/mL). However, the impurity rejection for both impurities 18 and 19 using a pure
organic solvent was minimal. Introduction of water as a coꢀsolvent enabled significant
reduction of both impurities.32 Since over 20% mass loss was observed with 5% of water
in MEK, 3ꢀ4% of water in MEK was chosen for the salt formation step to balance the
recovery of API and the purge of impurities. GDCꢀ0994•benzenesulfonate salt was
isolated in 71% yield over three steps and in 99.6 A% HPLC and >99.5% ee 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
Scheme 9. One-pot desilylation and benzenesulfonate salt formation
SO3H
N
N
O
Cl
F
O
Cl
F
HN
N
HN
N
N
N
N
N
THF/water
N
N
9
BsOH
GDC-0994
OTBS
OH
82% over three steps from 7
99.6% purity, >99.5% ee
individual impurity <0.25%
One-pot Desilylation and Benzenesulfonate Salt Formation using
Benzenesulfonic Acid. While the end game process via the telescoped deprotection / salt
formation provided GDCꢀ0994 API efficiently, we realized that the whole process could
be further streamlined by the removal of unnecessary unit operations. Since desilylation
only required a catalytic amount of acid, we envisioned that benzenesulfonic acid could
serve a dual purpose as a catalyst for deprotection and acid for salt formation. After acidꢀ
catalyzed desilylation, the crystalline benzenesulfonate salt would precipitate out of the
reaction solution. As a result, a freebasing step and lengthy solvent switches could be
removed.
Although the genotoxic impurity methyl benzenesulfonate was well controlled and
ACS Paragon Plus Environment

Page 23 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
closely monitored in the telescoped process, it is ideal to avoid alcoholic solvents in the
salt formation step and completely eliminate genotoxic impurity concerns. After the SNAr
reaction in the optimal THF / 2ꢀMeTHF solution was completed, the solvents were
swapped to THF / water for the subsequent desilylation and salt formation (Scheme 9).
Our initial attempts to perform a crystallizationꢀdriven reaction to form the API
benzenesulfonate salt in THF with a lower water content (<3 wt%) were successful but
resulted in an uncontrolled precipitation of the product. This led to unreacted
intermediate 9 being trapped in the solids and a high level of 9 (1ꢀ2A% HPLC) in the
crude API. To alleviate this issue we decided to perform the deprotection reaction in a
homogeneous solution with a higher water content in THF (10 wt%). After complete
deprotection, excess water was removed through azeotropic distillation. The solvent
composition was then adjusted into the metastable zone before seeds were introduced for
a controlled crystallization. As a result, the API was obtained with 82% isolated yield
over 3 steps and excellent purity (>99.5A% HPLC and >99.5% ee, no individual
impurities were observed over 0.25%).
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
Recrystallization of crude API. Although the purity of crude API was high enough
to meet our specifications, a recrystallization was developed and implemented to further
ensure API quality and a controlled unimodal distribution of particle size [d (0.5) < 50
micron]. The crude GDCꢀ0994•benzenesulfonate salt was dissolved at 78 °C in EtOH
after slowly introducing water to ~10 wt%. After a hot polish filtration, seeds were
charged at 60 °C to induce crystallization. The water content of the solution was then
adjusted to 2ꢀ3 wt% via an azeotropic distillation at 60 °C to maximize recovery. The
recrystallization offered decent purging of impurities 17 (>10×), 18 (2×) and 19 (3×).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 40
1
2
3
4
GDCꢀ0994•benzenesulfonate API was isolated in 89% yield.
5
6
3. CONCLUSION
7
8
9
In summary, an efficient and scalable synthetic route to the ERK inhibitor, GDCꢀ
0994•benzenesulfonate was developed and demonstrated on a multiple kilogram scale.
The API was prepared in 7 steps, 41% overall yield and high purity of >99.5A% HPLC
and >99.5% ee (Scheme 10). A highly enantioselective enzymatic ketone reduction was
developed as an efficient approach to the chiral diol intermediate. The performance and
regioselectivity of the pyridone SN2 reaction was improved from ~50% yield (N vs O
ratio 80:20) to 74% yield (N vs O ratio, 89:11) with careful selection of base and solvent
screens. The sulfide oxidation with mꢀCPBA was replaced by green and efficient
tungstateꢀcatalyzed oxidation with hydrogen peroxide. The reaction safety of the
oxidation was evaluated to achieve a safe and robust process. The SNAr displacement
reaction was optimized to control levels of impurities observed in the final API. The end
game steps were streamlined with development of a telescoped SNAr / desilylation /
benzenesulfonate salt formation process. Finally, development of the API
recrystallization enabled further purging 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
ACS Paragon Plus Environment

Page 25 of 40
Organic Process Research & Development
1
2
3
4
Scheme 10. Practical Synthesis of GDC-0994
5
6
7
8
N
O
1wt% KRED/cofactor
1.04 equiv D-glucose
9 wt% [sub]
Cl
F
Cl
F
Cl
F
MeS
N
1. DIPEA/DMAP,
TBSCl in DCM
NH
HO
HO
O
MsO
5
2. MsCl
9
pH 6.5 buffer
30 °C, 24 h
100%
KHMDS
dioxane/NMP
90 °C, 6 h
HO
11
TBSO
2
4
87% over 2
steps from 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
99.5% ee
74%
NH2
30% H2O2,
N
N
Na2WO4 (1 mol%)
n-PrOH, water
2,6-lutidine
N
N
N
O
Cl
F
O
Cl
O
Cl
F
MeS
N
S
N
HN
N
O
8
N
O
N
N
80 °C, 2 h
87%
F
N
N
6
LiHMDS
2-MeTHF/THF
-25 °C
7
OTBS
OTBS
9
OTBS
SO3H
N
N
O
Cl
F
O
Cl
F
HN
N
HN
N
N
N
EtOH/water
89%
N
N
N
N
THF/water
82% over three
steps from 7
OH
BsOH
GDC-0994
OH
BsOH
GDC-0994
>99.5%purity, 99.5% ee
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 40
1
2
3
4
4. EXPERIMENTAL SECTION
5
6
7
8
General. All commercially available reagents and solvents, including anhydrous
solvents, were used without further purification. Assay yields were obtained using
analytical standards prepared by recrystallization or preparative chromatography. All
isolated yields reflect correction for purity based on HPLC analyses or qNMR assays. 1H
(400 MHz), 13C (100 MHz), and 19F (282 MHz) NMR spectra were recorded at room
temperature (RT) on a Bruker spectrometer. Chemical shifts (δ) are expressed in parts per
million (ppm) relative to the corresponding deuterated solvent peak. HPLC analyses were
conducted on an Agilent 1200 series (Agilent Zorbax SBꢀPhenyl column (150 x 4.6 mm,
3.5 ꢁm); and mobileꢀphase gradients consisting of 0.05% TFA in water and 0.05% TFA
in acetonitrile) or (Waters Xbridge Phenyl column (150 x 4.6 mm, 3.5 ꢁm); and mobileꢀ
phase gradients consisting of 0.05% TFA in water and 0.05% TFA in acetonitrile). Chiral
HPLC analyses were conducted on an Agilent 1200 series (Chiralpak IA column (250 x
4.6 mm, 5 ꢁm); isocratic mobile phase 80% nꢀhexane and 20% ethanol). Analytical
characterizations were obtained on materials that were representative or purified for that
purpose.
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
(R)-1-(4-chloro-3-fluorophenyl)ethane-1,2-diol (2). To a 500 L reactor under N2
was charged water (150 kg), 4ꢀmorpholine ethanesulfonic acid (0.90 kg, 4.6 mol) and
anhydrous magnesium chloride (0.030 kg, 0.3 mol). After a clear solution was obtained,
nꢀheptane (37 kg), 11 (30 kg, 159 mol), Dꢀ(+)ꢀglucose monohydrate (34.8 kg, 176 mol)
and PEG 6000 (30.0 kg) were charged. The pH of the solution was adjusted to 6.5ꢀ7.0 by
a 1N aqueous NaOH solution at 28ꢀ32 °C. To the resulting solution was charged GDHꢀ
105 (0.30 kg), βꢀdiphosphopyridine nucleotide (0.30 kg, 0.45 mol) and KREDꢀNADHꢀ
ACS Paragon Plus Environment

Page 27 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
112 (0.30 kg). The resulting suspension was stirred at 29ꢀ31 °C for 10ꢀ12 h while
maintaining the pH of the reaction mixture at 6.5ꢀ7.0 by the controlled addition of a 1N
aqueous NaOH (160 kg). The pH of the reaction mixture was adjusted to 1ꢀ2 by the
addition of 49% H2SO4 (20 kg) to quench the reaction. EtOAc (271 kg) was then added
and the mixture was stirred at 20ꢀ30 °C for 10ꢀ15 min. The mixture was filtered through a
pad of celite and the cake was washed with EtOAc (122 kg). The combined organic
layers were separated and aqueous layer was extracted with EtOAc (150 kg). The organic
layers were combined and washed with water (237 kg). The pH of the mixture was
adjusted to 7.0ꢀ8.0 by addition of solid NaHCO3. The organic layer was separated,
concentrated and then switched to CH2Cl2. The crude product was obtained as a CH2Cl2
solution (30.9 kg, 100% assay yield, 99A% HPLC purity and >99.5% ee, 7.9 wt% in
CH2Cl2).
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
(R)-2-((tert-butyldimethylsilyl)oxy)-1-(4-chloro-3-fluorophenyl)ethyl methanesulfonate
(4). To a 1000 L reactor under nitrogen was charged with 2 (30.9 kg in 390 kg of
anhydrous CH2Cl2, 163 mol). The solution was cooled to –5 °C. Tertꢀ
butylchlorodimethyl silane (26.9 kg, 173 mol) was added in portions at –5 °C. A solution
of 4ꢀdimethylaminopyridine (0.995 kg, 3.3 mol) and triethylamine (43 kg, 426 mol) in
anhydrous CH2Cl2 (135 kg) was added dropwise to the above solution at –5 °C. The
reaction solution was warmed to 25 °C over 2.5 h and stirred for 12 h. The obtained
solution was cooled to –5 °C. A solution of methanesulfonyl chloride (20.4 kg, 179 mol)
in anhydrous CH2Cl2 (135 kg) was added to the above solution over 4 h and the
temperature was controlled below 0 °C. After stirring at 5 °C for 2 h, the reaction was
washed with water (229 kg), followed by 5 wt% aqueous citric acid (236 kg), 2 wt%
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 40
1
2
3
4
5
6
7
8
9
NaHCO3 aqueous solution (228 kg) and finally water (228 kg). The resulting CH2Cl2
solution was dried over anhydrous Na2SO4 (1.5 kg), filtered, rinsed with CH2Cl2 (38 kg)
and concentrated in vacuo below 15 °C to afford crude 4 as a CH2Cl2 solution (54.8 kg in
206 kg of CH2Cl2, 94.9A% HPLC purity, 26.6 wt% and 88.0% assay yield, KF = 0.01%).
(S)-1-(2-((tert-butyldimethylsilyl)oxy)-1-(4-chloro-3-fluorophenyl)ethyl)-4-(2-
(methylthio)pyrimidin-4-yl)pyridin-2(1H)-one (6). In a 1000 L reactor, a solution of 4
(197.9 kg, 26.6% assay, 130 mol) in CH2Cl2 was switched to THF (301 kg) to remove
residual CH2Cl2 (KF = 0.01% and residual CH2Cl2 = 1%). The resulting solution was then
concentrated to 3 volumes and transferred to a drum. To a 1000 L reactor was charged 5
(27.1 kg, 96.1 wt% assay solid, 119 mol), anhydrous NMP (14 kg) and anhydrous 1,4ꢀ
dioxane (417 kg). A solution of KHMDS (124 kg, 1M in THF) was added dropꢀwise
below 40 °C under N2. The mixture was stirred at 25ꢀ40 °C for 3 h. The pretreated
solution of 4 in THF was added while maintaining the temperature below 40 °C. The
reaction solution was warmed to 80ꢀ90 °C and stirred at the same temperature for 6 h.
The reaction mixture was concentrated to 70 L to remove dioxane at 40 °C in vacuo. To
the reaction mixture was added 547 kg of CH2Cl2 and 160 kg of water at 15 °C. The pH
of the solution was adjusted to 6.0 with 30 wt% citric acid aqueous solution (6.4 kg).
After the phase separation, the CH2Cl2 layer and middle emulsion layer were filtered
through a celite (9 kg) pad. The organic layer was separated and washed with 3 wt%
Na2SO4 (2 x 214 kg). The combined aqueous layers were backꢀextracted with CH2Cl2
(168 kg). The combined organic layers were filtered through a silica gel pad (15 kg) and
the pad was washed with CH2Cl2 (134 kg). The combined filtrates were concentrated to
80 L. The solution was then switched with nꢀheptane in three portions (180 kg + 84 kg +
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 29 of 40
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
134 kg) to remove CH2Cl2 and concentrated to 80 L (2 volumes) at 30 °C under reduced
pressure. The slurry was cooled to 0ꢀ10 °C and stirred at the same temperature for 10 h.
The solid was isolated and dried at 35ꢀ45 °C for 16 h to afford 6 (49.6 kg, 74% isolated
yield, 93.9 wt% NMR assay, 95.5A% HPLC purity and >99.5% ee) as a brown solid. mp
124.6 °C; 1H NMR (400 MHz, CDCl3): 8.64 (d, J = 5.2 Hz, 1H), 7.44 (d, J = 7.6 Hz, 1H),
7.41 (t, J = 8.0 Hz, 1H), 7.33 (d, J = 5.2Hz, 1H), 7.30ꢀ7.27 (m, 2H), 7.14 (d, J = 8.4 Hz,
1H), 6.84 (dd, J = 7.6, 2.0 Hz, 1H), 6.22 (t, J = 4.0 Hz, 1H), 4.33 (dd, J = 11.6, 4.0 Hz,
1H), 4.21 (dd, J = 11.6, 4.0 Hz, 1H), 2.63 (s, 3H), 0.86 (s, 9H), 0.039 (s, 3H), ꢀ0.02(s,
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
13
3H). C NMR (100 MHz, CDCl3): 173.57, 162.52, 160.90, 158.43, 158.24 (d, J = 248
Hz), 147.04, 138.31 (d, J = 6.5 Hz), 137.20, 131.03, 125.07 (d, J = 3.6 Hz), 120.93 (d, J
= 17.5 Hz), 118.53, 117.0 (d, J = 22.1 Hz), 112.44, 103.28, 62.63, 57.88, 25.81, 18.17,
14.33, ꢀ5.62, ꢀ5.66. 19F NMR (282 MHz, CDCl3) ꢀ114.15 (dd, J =10.0, 7.6). HRMS calcd
for C24H29ClFN3O2SSi [M + H] 506.1501, found 506.1490.
(S)-1-(2-((tert-butyldimethylsilyl)oxy)-1-(4-chloro-3-fluorophenyl)ethyl)-4-
(2-(methylsulfonyl) pyrimidin-4-yl)pyridin-2(1H)-one (7): Oxidation with m-CPBA.
To a 1000 L reactor was charged 6 (48.1 kg, 93.9 wt%, 90 mol) and CH2Cl2 (664 kg). To
the resulting solution was charged mꢀCPBA (39 kg, 169 mol, 75 wt%) in portions while
maintaining the temperature below 10 °C. The reaction was stirred at 5ꢀ10 °C for 14 h.
The reaction mixture was transferred to a second reactor containing a 6 wt% aqueous
NaHCO3 solution (620 kg) over 1.5 h at 10ꢀ20 °C. The organic layer was separated and
the aqueous layer was extracted with CH2Cl2 (73 kg). The combined organic layers were
added into a 4.5 wt% aqueous NaHCO3 solution (410 kg). To the resulting mixture at 15ꢀ
20 °C was charged a 9 wt% aqueous Na2S2O3 solution (100 kg) dropꢀwise to consume
ACS Paragon Plus Environment