Improvements to Enable the Large Scale Synthesis of 1-{[(2S,3S,4S)-3-Ethyl-4-fluoro-5-oxopyrrolidin-2-yl]methoxy}-7-methoxyisoquinoline-6-carboxamide (PF-06650833)

Article abstract of DOI:10.1021/acs.oprd.8b00386

An improved process for the large scale synthesis of 1-{[(2S,3S,4S)-3-ethyl-4-fluoro-5-oxopyrrolidin-2-yl]methoxy}-7-methoxyisoquinoline-6-carboxamide (1), a candidate currently in clinical development, was developed. Key objectives were to eliminate chro

Full text of DOI:10.1021/acs.oprd.8b00386

Subscriber access provided by Kaohsiung Medical University
Full Paper
Improvements to Enable the Large Scale Synthesis of 1-
{[(2S,3S,4S)-3-Ethyl-4-fluoro-5-oxopyrrolidin-2-yl]methoxy}-7-
methoxyisoquinoline-6-carboxamide (PF-06650833)
Stephen W Wright, Bryan Li, Zhihui Peng, Lulin Wei, Emma L
McInturff, David Place, David B Damon, and Robert A. Singer
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.8b00386 • Publication Date (Web): 27 Nov 2018
Downloaded from http://pubs.acs.org on November 27, 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 46
Organic Process Research & Development
1
2
3
4
5
6
Improvements to Enable the Large Scale Synthesis of 1-{[(2S,3S,4S)-3-Ethyl-4-fluoro-5-oxopyrrolidin-2-
yl]methoxy}-7-methoxyisoquinoline-6-carboxamide (PF-06650833)
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
Stephen W. Wright*,§ Bryan Li,ǂ Zhihui Peng,ǂ Lulin Wei,ǂ Emma McInturff,ǂ David Place,ǂ David B.
Damon,ǂ Robert A. Singerǂ
§Medicine Design, Pfizer Worldwide Research and Development, 445 Eastern Point Road, Groton,
Connecticut 06340, United States
ǂChemical Research and Development, Pfizer Worldwide Research and Development, 445 Eastern Point
Road, Groton, Connecticut 06340, United States
Email: stephen.w.wright@pfizer.com
TOC Graphic:
O
F
HN
O
O
O
O
F
S
HN
N
N
O
O
O
MeO
HO
N
7
11
276 kg
2
H2NOC
165 kg
100 kg
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 46
1
2
3
4
Keywords: diastereoselective, fluorination, cuprate, conjugate addition, sulfoxide, elimination, γ-lactam
5
6
7
8
9
Abstract:
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
An improved process for the large scale synthesis of 1-{[(2S,3S,4S)-3-ethyl-4-fluoro-5-
oxopyrrolidin-2-yl]methoxy}-7-methoxyisoquinoline-6-carboxamide (1), a candidate currently in clinical
development, was developed. Key objectives were to eliminate chromatographic purifications, to
maximize the reproducibility of each step, and to improve the yield and efficiency of each step relative to
the previous discovery syntheses of 1. This work was focused on improvements to the synthesis of the
stereochemically complex lactam 2. Steps of particular concern were the preparation of the unsaturated
lactam 6, the cuprate conjugate addition reaction to produce 7, and the conversion of 7 to 8 with a high
degree of diastereoselection. The solutions to these challenges have permitted the synthesis of 2 in excess
of 100 kg, which in turn has permitted 1 to be prepared in sufficient amounts to support further
development.
ACS Paragon Plus Environment

Page 3 of 46
Organic Process Research & Development
1
2
3
4
Introduction:
5
6
7
8
9
Interleukin-1 receptor associated kinase 4 (IRAK-4) is a serine threonine kinases that plays a key
role in innate immune signaling. IRAK-4 is activated by the interleukin (IL-1) family receptors (IL-1R,
IL-18R, and IL-33R), as well as the Toll-like receptors (TLRs). Inhibition of IRAK-4 blocks the
production of inflammatory cytokines such as type I interferons, tumor necrosis factor (TNF), IL-1, IL-6,
and IL-12 that are key drivers of autoimmune and inflammatory diseases. IRAK-4 is an attractive
therapeutic target for diseases associated with dysregulated inflammation, such as systemic lupus
erythematosus and rheumatoid arthritis.1 Recently, we described the synthesis and biological activity of 1-
{[(2S,3S,4S)-3-ethyl-4-fluoro-5-oxopyrrolidin-2-yl]methoxy}-7-methoxyisoquinoline-6-carboxamide
(PF-06650833, 1).2 The principal challenges with the synthesis of 1 arise from the need to control the
stereochemistry of the lactam fragment 2, which contains three contiguous stereocenters, the integrity of
which must be maintained throughout the synthesis (Scheme 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1: Retrosynthesis of 1
O
F
HN
O
Cl
MeO
NC
F
+
HN
N
O
MeO
HO
N
2
3
H2NOC
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 46
1
2
3
4
5
6
7
8
9
The stereocenter at C-5 is derived from the chiral pool, starting with (S)-pyroglutamic acid. The
stereocenter at C-4 is then introduced via a diastereoselective cuprate conjugate addition reaction. Lastly,
the stereocenter at C-3 is introduced by enolate formation and reaction with an appropriate electrophile.
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
Results and Discussion:
The original synthesis of 1 was carried out as shown in Scheme 2, starting with the known keto-
aminal of (S)-pyroglutaminol 4.3 The olefin was introduced in 73% yield by a modified Tsuji oxidation4
of the trimethylsilyl enol ether 5, rather than the more customary approach via a selenoxide elimination.5
The resulting unsaturated lactam 6 was treated with an excess (1.5 equivalents) of an organocopper
reagent prepared from equimolar amounts of lithium (2-thienyl)cyanocuprate6 (LiThCN) and
ethylmagnesium chloride in the presence of 2.0 equivalents of chlorotrimethylsilane (TMSCl) at -70 C to
afford the ethyl substituted lactam 7 in 90% yield with high (30:1) syn diastereoselection.7 Certain ,β-
unsaturated γ-lactam substrates have been shown to undergo productive conjugate addition reactions with
organocopper reagents in the absence of TMSCl.8 However, the addition of TMSCl has been found to be
necessary for productive reaction of N-alkyl ,β-unsaturated γ-lactams, which otherwise suffer γ-
deprotonation.9 The combination of organocopper reagents with TMSCl has been found to accelerate and
ACS Paragon Plus Environment

Page 5 of 46
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
improve conjugate additions,10 and may accelerate the collapse of the d-π complex with concurrent C-C
bond formation.11 Thus the TMSCl mediated conjugate addition of organocopper reagents to N-alkyl ,β-
unsaturated γ-lactams has been well established.12 These reactions must be conducted at temperatures
below -55 C to prevent alkylation of the TMSCl by the organocopper reagent.10 The conjugate addition
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
of organocopper reagents to 6 was previously unknown and occurred with a reversal of the
diastereoselection observed previously.7, 12
Scheme 2: First Discovery Synthesis of 1a
SiMe3
O
O
O
O
O
O
a
b
c
d
F
F
+
N
N
N
N
N
N
O
O
O
O
O
O
F
4
6
7
8
5
9
O
O
F
F
HN
O
O
HN
O
Cl
f
e
g
MeO
NC
+
HN
N
8
MeO
MeO
NC
HO
N
N
2
3
H2NOC
10
1
aConditions: (a) LDA (1.2 equiv), TMSCl (1.3 equiv), THF, -60 C, 30 min; (b) allyl methyl carbonate
(1.1 equiv), Pd(OAc)2 (0.05 equiv), THF, 65 C, 2 h, 73% (2 steps); (c) LiThCN (1.5 equiv), EtMgCl (1.5
equiv), TMSCl (2.0 equiv), THF, -78 C, 6 h, 90%; (d) LDA (1.8 equiv), NFSI (1.25 equiv), THF, -78
C, 1 h, 23% (8), 45% (9); (e) pTsOH (0.05 equiv), MeCN, H2O, 90 C, 2 h, 97%; (f) 3 (0.9 equiv),
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 46
1
2
3
4
5
6
KHMDS (2.0 equiv), DMF, THF, -10 C, 30 min, 84%; (g) H2O2 (10 equiv), K2CO3 (4.0 equiv), DMSO,
20 C, 2 h, 97%.
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
Fluorination to afford 8 was accomplished in 23% yield by enolate formation with 1.8 equivalents of
LDA at -70 C followed by treatment with 1.25 equivalents of N-fluoro(bis(benzenesulfonyl)imide)
(NFSI), a well established reagent for the fluorination of γ-lactam enolates.13 In this reaction,
diastereoselection was relatively poor (2:1) and favored the undesired anti isomer 9 (45% yield).
Nevertheless, sufficient 8 could be obtained in this way to support medicinal chemistry efforts. Cleavage
of the acetonide protecting group under acidic conditions (catalytic pTsOH in aqueous MeCN at 90 C)
yielded the alcohol 2 in 97% yield following purification by chromatography. The synthesis was
completed by coupling the alcohol 2 to the chloroisoquinoline 314 via a SNAr reaction using 2.0
equivalents of potassium hexamethyldisilazide (KHMDS)15 in a mixture THF and DMF at -10 C to
afford the intermediate 10 in 84% yield.16 No epimerization of the fluoro group was observed under these
conditions.17 Nitrile hydrolysis using 10 equivalents of H2O2 and 4.0 equivalents of K2CO3 in DMSO18 at
20 C completed the synthesis and provided 1 in 97% yield following purification by chromatography.
ACS Paragon Plus Environment

Page 7 of 46
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Four improvements were made in order to facilitate the synthesis of larger amounts of 1 to
support in vivo and preclinical safety studies (Scheme 3).14 First, the organocopper reagent was prepared
from copper(I) bromide – dimethyl sulfide complex19 (2.0 equivalents) and ethylmagnesium bromide (4.0
equivalents) in THF at -5 C. This allowed us to move away from the relatively dilute conditions of 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
LiThCN chemistry and eliminated some additional problems associated with that reagent: the need to
generate the LiThCN reagent, to remove thiophene from the crude product, and to manage a cyanide
containing waste stream. This provided the ethyl lactam 7 in 88% yield with the same high
diastereoselection observed previously. Second, the unwanted anti fluorination isomer 9 was epimerized
to provide more of the desired all syn isomer 8 by means of a kinetic protonation (Scheme 4). The anti
fluoro isomer 9 was converted to the enolate 11 with 1.1 equivalents of LDA at -70 C in toluene,
followed by protonation with 2.0 equivalents of a sterically hindered acid (2,6-di-t-butyl-4-methylphenol,
BHT) at a temperature below -65 C. This procedure afforded a 44% yield of 8 and 27% of the anti
isomer 9 following chromatography. This procedure improved the throughput of the desired isomer 8,
albeit at the cost of an added step and further silica gel chromatography.20 Third, the acetonide was
cleaved with catalytic TFA in aqueous MeCN at 65 C. This allowed the lactam alcohol 2 to be isolated
in 98% yield by concentration followed by crystallization and drying without need for chromatography.
Lastly, nitrile hydrolysis of 10 was effected using MsOH as reagent and solvent at 60 C instead of using
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 46
1
2
3
4
the K2CO3 - H2O2 - DMSO conditions used previously, which were perceived to be less desirable for
5
6
7
scale up.21 The crude 1 was isolated by extraction and purified by recrystallization to provide 1 in 91%
8
9
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
Scheme 3: First Generation Scale Up Synthesis of 1a
O
O
O
O
a
b
F
F
+
N
N
N
N
O
O
O
O
6
7
8
9
O
O
F
F
HN
O
HN
O
Cl
e
c
d
MeO
NC
F
+
HN
N
8
O
MeO
H2NOC
MeO
NC
HO
N
N
2
3
10
1
aConditions: (a) CuBr  Me2S (2.0 equiv), EtMgBr (4.0 equiv), TMSCl (1.5 equiv), Et2O, THF, -70 C, 4
h, 88%; (b) LDA (1.05 equiv), NFSI (1.05 equiv), THF, -70 C, 2 h, 37% (8), 54% (9); (c) TFA (0.2
equiv), MeCN, H2O, 65 C, 4 h, 98%; (d) 3 (0.9 equiv), KHMDS (2.0 equiv), DMF, THF, -10 C, 75 min,
90%; (e) MsOH (7.5 equiv), 60 C, 26 h, 91%.
Scheme 4: Epimerization of 9 to 8 by Kinetic Protonationa
ACS Paragon Plus Environment

Page 9 of 46
Organic Process Research & Development
1
2
3
4
5
6
7
8
O
OLi
11
O
8
O
a
b
F
F
F
F
+
N
N
N
N
O
O
O
O
9
9
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
aConditions: LDA (1.1 equiv), PhCH3, -78 C, 2.5 h; (b) 2,6-di-t-butyl-4-methylphenol (2.0 equiv), -65
C, 3.5 h, 44% (8), 27% (9).
In order to provide sufficient 1 for further development, a significant route optimization effort
was undertaken. Key objectives of this work were to maximize the reproducibility of each step, eliminate
the chromatographic purification of intermediates, improve the yield and efficiency of each step, and
reduce raw material costs. Three transformations warranted especial attention:
(1) The preparation of the unsaturated lactam 6 via the Tsuji oxidation was difficult to carry out
on larger scales, due to the need to obtain the intermediate silyl enol ether 5 free from both
diisopropylamine and lithium chloride.22 This required distillation of the diisopropylamine and THF
solvent with hexanes followed by filtration of a hexane solution of 5 to remove the lithium chloride.
Failure to completely remove these intermediate by-products would result in stalling of the Tsuji catalytic
cycle and recovery of the starting lactam 4.
(2) The cuprate conjugate addition reaction faced two major hurdles to further scale up. The
greatest concern was the preparation the organocopper reagent at temperatures between -3 and -30 C,
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 46
1
2
3
4
followed by cooling of the resulting organocopper solution to -70 C prior to the sequential addition of
5
6
7
8
9
TMSCl followed by lactam 6. This was problematic due to the amount of time required to accomplish
both the preparation of the organocopper reagent and also the subsequent cooling operation, time periods
that necessarily increased with increasing scales. The organocopper solution was found to degrade with
time and therefore decomposition of the reagent became exacerbated at larger scales due to the increased
times required to prepare and cool larger volumes. Lacking an in-process control procedure to reliably
determine the potency of the organocopper reagent, this was viewed as a serious risk for failure of this
step. In addition, the use of stochiometric or super-stochiometric amounts of copper salts would result in
unacceptably large waste streams. The use of large amounts of CuBr  Me2S in particular would have
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
been prohibitively expensive and have required scrubbing procedures to trap the evolved Me2S.
(3) The most significant problem was clearly the fluorination of 7 to produce the desired all syn
lactam 8. Further use of NFSI to accomplish this transformation would have resulted in unacceptable
reagent costs and could have exceeded supply capacity. More importantly, the use of NFSI provided the
desired isomer 8 as the minor diastereomer and the diasterereomeric ratio was not materially influenced
by selection of reaction conditions. While the BHT mediated epimerization procedure (vide supra) was
successfully implemented in early synthetic work, both the NFSI fluorination and epimerization
procedures required extensive chromatographic purifications which were not deemed acceptable for
ACS Paragon Plus Environment

Page 11 of 46
Organic Process Research & Development
1
2
3
4
larger scale work. Initially a dynamic kinetic epimerization by crystallization procedure was investigated,
5
6
7
which capitalized on the fact that the desired 8 is a solid (mp 55 - 57 C) while the undesired isomer 9 is
8
9
an oil at 20 C. However, this procedure proved to be difficult to scale up and was better suited to
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
upgrading material that contained primarily 8 than it was to the conversion of large excesses of 9 to 8.
With these limitations in mind, research was conducted to identify appropriate solutions to permit
the synthesis of 1 to be conducted successfully in amounts of thirty kilograms or more. As a result of this
effort, solutions to these principal issues were identified and further optimization of other steps was
accomplished to facilitate the purification and isolation of intermediates and eliminate chromatography.
The solutions to the three principal issues noted previously were the following:
(1) The thermal elimination of the previously unknown sulfoxide 12 provided convenient
introduction of the olefin to access unsaturated lactam 6 (Scheme 5). The sulfoxide was readily prepared
from 4 by a base mediated condensation with 1.5 equivalents of methyl benzenesulfinate23 using NaH24 in
THF at around 45 C to afford the sulfoxide 12 after quenching with water, acidification with AcOH,
extraction, and crystallization from a mixture of heptane and MTBE to afford the sulfoxide 12 in 72%
yield as an equal mixture of two diastereomers as determined by 1H NMR.25 Formation of sulfoxides such
as 12 in one step by enolate condensation has been little used in γ-lactams,26 with two step procedures
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 46
1
2
3
4
5
6
7
8
9
involving sulfide formation and subsequent oxidation being more common.27 The one step condensation
procedure which avoids the need for an oxidation step was preferred. While it might be anticipated that
PhSO2Me could react to transfer a methyl group, no Me transfer products were observed. Elimination of
12 to provide 6 was accomplished by heating a toluene solution of 12 at 110 C28 in the presence of 3.0
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
equivalents of anhydrous disodium phosphate29 until the elimination was completed, after which the
mixture was cooled, diluted with heptane, and 6 was extracted into a 1:6 (v/v) mixture of methanol and
water.30 The aqueous solution was then extracted with CH2Cl2 with the addition of NaCl to facilitate
phase separation.31 The CH2Cl2 extract was dried with Na2SO4 and partially concentrated prior to solvent
exchange with THF. The resulting THF solution of 6 was taken into the next step as – is.
Scheme 5: Preparation of 7 by Telescoped Sulfoxide Elimination and Cuprate Additiona
O
O
O
O
O
a
b
c
S
N
N
N
N
O
O
O
O
4
12
6
7
aConditions: (a) PhSO2Me (1.5 equiv), NaH (2.1 equiv), THF, 45 C, 11 h, 72%; (b) Na2HPO4 (3.0
equiv), PhMe, 110 C, 14 h; (c) EtMgBr (1.2 equiv), TMSCl (1.1 equiv), CuI (0.07 equiv), -70 C, 4.5 h,
87% (2 steps).
ACS Paragon Plus Environment

Page 13 of 46
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
(2) The keys to successful scale up of the cuprate addition were the discoveries that this reaction
could be carried out with catalytic amounts of copper(I) and that CuI was entirely satisfactory for this
purpose.32 While the use of catalytic amounts of copper(I) salts has been reported for the conjugate
addition of Grignard reagents to N-sulfonyl ,β-unsaturated γ-lactams,33 the use of catalytic amounts of
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
copper(I) salts to effect the conjugate addition of Grignard reagents in the presence of TMSCl has not
been widely practiced.34 This is likely due to concern about the potential of TMSCl to silylate the
Grignard reagent directly to afford an organyltrimethylsilane, and in small scale work there is little
incentive to investigate the use of catalytic amounts of copper(I) salts. However, a review of the literature
suggested that the direct alkylation of TMSCl by Grignard reagents was rather slow at temperatures of 0
C or below.35 Thus we had reason to anticipate that a mixture of EtMgBr, TMSCl, and catalytic Cu(I)
could be maintained at < -60 C without consumption of the TMSCl by either the EtMgBr or the
organocopper species present. Experimentation showed that indeed a mixture of excess EtMgBr and
catalytic (<10 mole percent) CuI could be prepared at -60 C and treated with TMSCl at -70 C without
significant reaction between the TMSCl with the Grignard reagent, which would have prevented a
successful conjugate addition reaction upon addition of 6. This procedure eliminated the time dependent
cuprate formation and cooling steps, reduced the amount of TMSCl and Grignard regent employed, and
greatly reduced the amount of copper salts required for successful reaction while still delivering 7 with
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 46
1
2
3
4
5
6
7
8
9
excellent yields and diastereoselectivity. The substitution of the more readily available CuI for CuBr 
Me2S reduced costs and eliminated the need to control Me2S with no effect on reaction profile. As
performed, a suspension of 7 mole percent36 of CuI in THF was cooled to -60 C and treated with 1.2
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
equivalents of 1 M EtMgBr in THF. The mixture was cooled to -70 C and TMSCl (1.1 equivalents) was
added, followed by the THF solution of 6 prepared previously. The mixture was kept at -70 C until 6 had
been consumed, after which the reaction mixture was transferred into a solution of NH4Cl. Following
completion of the quench, the mixture was filtered and extracted with MTBE. The extract was washed
with brine, filtered, and solvent exchanged with THF to provide a THF solution of 7 in 87% yield and
94% purity with no detectable amounts of other stereoisomers of 7 present.
(3) Among the options that could be considered to convert 7 to 8, the introduction of an alcohol
function anti to the ethyl group, followed by displacement with inversion to afford the syn fluoride was
considered the most desirable (Scheme 6). A significant cost savings could be realized by changing from
an electrophilic to a nucleophilic fluorination reagent if (a) a highly diastereoselective synthesis of the
alcohol 14 could be realized, and (b) deoxofluorination of 14 could be accomplished with excellent
preservation of stereochemical integrity. Previous work during the medicinal chemistry effort had shown
that diastereoselective anti functionalization of the enolate 13 could be achieved with some reagents,
including anti selective hydroxylation with bulky oxaziridine reagents.37 A significant effort was therefore
ACS Paragon Plus Environment

Page 15 of 46
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
put into the identification of a highly diastereoselective enolate oxidation procedure, one which could be
carried out safely and which employed an inexpensive, readily available oxidizing agent.38 Initial work
using oxygen at -70 C to oxidize the lithium enolate 13 gave a 4:1 ratio of diastereomers in favor of 14.
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
However, the use of lithium t-butylperoxide (t-BuOOLi) afforded 14 with greater than 30:1
diastereoselectivity. To be successful, this reaction required anhydrous t-butyl hydroperoxide which was
recognized to be a potential safety hazard on large scale.39 Accordingly, the reaction sequence was
adapted to a flow process that simultaneously generated an anhydrous solution of t-butyl hydroperoxide in
nonane in situ using a membrane pervaporation procedure, and consumed this solution by reaction with
the lithium enolate 13. The ultimate result of that work was a flow process that accomplished (1) the
drying of the t-butyl hydroperoxide feedstock, (2) the formation of the lithium enolate 13 from 7 using
lithium hexamethyldisilazide (LiHMDS), and (3) the subsequent oxidation of 13 to afford 14 in 74%
yield, as a solution that was carried directly to the fluorination step.
Scheme 6: Diastereoselective Oxidation of 7 to 14a
O
O
OLi
a
b
OH
N
N
N
O
O
O
7
13
14
aConditions: (a) LHMDS (2.2 equiv), THF, -20 °C; (b) LiOOtBu, nonane, -20 °C → 0°C, 70% (2 steps).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 46
1
2
3
4
5
6
7
8
9
While many reagents are available to accomplish deoxofluorination, those formally derived from
SF4 are known to generally possess undesirable thermal stability profiles.40 Initial efforts to convert the
alcohol 14 to the fluoride 8 with inversion therefore converted the alcohol 14 to the triflate 15a with triflic
anhydride and 2,6-lutidine, which then underwent SN2 fluorination with triethylamine trihydrofluoride to
afford 8 in 60% yield.41 (Scheme 7). This procedure was an improvement over the original NFSI
chemistry with its attendant chromatography needs; however, the use of triethylamine trihydrofluoride as
a fluoride source and triflic ahydride, while manageable for large scale, were not considered ideal.42 The
use of inexpensive nonafluorobutanesulfonyl fluoride offered a more suitable option for the conversion of
14 to 8 on large scale. This reagent has seen steady acceptance as a deoxofluorinating reagent since its
introduction,43 most commonly with DBU or DIEA used as the base.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 7: Conversion of 14 to 7 with Inversion of Configurationa
O
O
O
a
b
F
OH
X
N
N
N
O
O
O
14
15
8
a,
X = OTf
X = OSO2C4F9
b,
ACS Paragon Plus Environment

Page 17 of 46
Organic Process Research & Development
1
2
3
4
5
6
aConditions: (a) Tf2O (1.2 equiv), 2,6-lutidine (1.5 equiv), CH2Cl2, -5 °C, 10 min (for 15a); C4F9SO2F
(1.3 equiv), TMEDA (1.4 equiv), PhMe, 20 C, 23 h (for 15b); (b) Et3N • 3HF (1.5 equiv), 2,6-lutidine
7
8
9
(3.0 equiv), 2-MeTHF, 20 C, 3 h (for 15a), 60% (2 steps); No further reagents (for 15b), 72%.
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 reagent permitted the alcohol activation and SN2 displacement by fluoride ion to be
accomplished in a single step, without need for the isolation of the intermediate sulfonate ester or the use
of an anhydrous, organic solvent - soluble fluoride ion source. Optimized conditions for the
transformation introduced the anti alcohol 14 as a solution in 2-MeTHF and THF from the flow oxidation
step. This was diluted with toluene and treated with TMEDA44 (1.4 equiv) followed by
nonafluorobutanesulfonyl fluoride (1.3 equiv) at 20 C for 16 h. Following water washing and extraction,
crystalline 8 could be isolated in 72% yield and 98% purity by the partial distillation of solvent, addition
of heptane and crystallization at 0 C.45
The remaining steps in the synthesis of 1 were the hydrolysis of the acetonide protecting group of
8, the SNAr reaction between 2 and 3 to afford 10, and final nitrile hydrolysis of 10 to afford 1. These
steps were carried out with relatively little modification from previous work, with the exception that
nitrile hydrolysis step was effected using H2SO4 instead of MsOH. The lactam alcohol 2 was isolated as a
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 46
1
2
3
4
5
6
7
8
9
crystalline solid from the acetonide hydrolysis. Optimized conditions for the acetonide cleavage used
TFA (0.2 equiv) in 3% aqueous MeCN, after which 2 was isolated by crystallization from i-PrOAc and
heptane in 79 to 90% yield. The SNAr reaction was conducted as described previously using 2.2
equivalents of KHMDS at -10 C, with the workup modified to permit isolation of the intermediate 10 as
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 crystalline solid. The reaction mixture was diluted with EtOAc, then quenched into a solution of
NaH2PO4 (4.0 equiv.) in water. The EtOAc was partially distilled, toluene was added and the mixture was
distilled further to remove EtOAc and crystallize 10 in 73% yield. The purified 10 was dissolved in
H2SO4 (19 equiv.) and heated to 80 C until nitrile hydrolysis was complete. The cooled mixture was
diluted into NH3 solution (47 equiv.) at 5 C, then crystallized by the addition of EtOH, warmed to
dissolve the solids present, and then allowed to cool to permit 1 to crystallize. The solid was filtered,
washed with water and dried prior to recrystallization from EtOH and water with decolorizing carbon to
afford the purified 1 in 82% yield.
Scheme 8. Conversion of 8 to 1a
O
O
F
F
HN
O
O
HN
O
Cl
a
b
MeO
NC
c
F
F
HN
+
N
N
O
O
MeO
MeO
NC
HO
N
N
8
2
3
H2NOC
10
1
ACS Paragon Plus Environment

Page 19 of 46
Organic Process Research & Development
1
2
3
4
aConditions: (a) TFA (0.2 equiv), H2O (6.7 equiv), MeCN, reflux, 5 h, 79%; (b) KHMDS (2.2 equiv),
5
6
7
THF, DMF, -10 C, 6.5 h, 73%; (c) H2SO4, 80 C, 4 h, 82%.
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
Summary
An improved process for the large scale synthesis of 1-{[(2S,3S,4S)-3-ethyl-4-fluoro-5-
oxopyrrolidin-2-yl]methoxy}-7-methoxyisoquinoline-6-carboxamide (1), a clinical candidate, was
developed. Key objectives for improvement of the discovery synthetic route were: (1) a scaleable
procedure for the synthesis of the unsaturated lactam 6; (2) an improved procedure to accomplish the
conjugate addition reaction to produce 7 in good yield and with high diastereoselectivity, with reduced
use of copper salts and no dependence upon cooling times; (3) the preparation of 8 from 7 in good yield
and with high diastereoselectivity; (4) the elimination of the intensive chromatography attendant to the
preparation and purification of 8, and (5) the reduction of costs associated with raw materials and
processing. A sulfoxide elimination provided a suitable method for the preparation of 6, while the
conjugate addition could be carried out with only catalytic amounts of CuI which eliminated the
troublesome dependence upon cooling times associated with this step. The preparation of 8 from 7 was
achieved by the highly diastereoselective oxidation of enolate 13 to 14 with t-BuOOLi in a flow process,
after which activation and SN2 inversion of 14 with nonafluorobutanesulfonyl fluoride provided 8.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 46
1
2
3
4
Successful scale up of the final three steps and purification by recrystallization allowed 1 to be prepared
5
6
7
in batches of greater than 30 kg at a time.46
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 new compounds were characterized by proton (1H) NMR spectra using Bruker spectrometers
and are reported in parts per million (ppm) relative to the residual resonances of the deuterated solvent.
Carbon (13C) NMR spectra (proton decoupled) were recorded similarly. Elemental Analyses were
performed by Intertek, 291 Rte. 22 East, PO Box 470, Whitehouse, NJ 08888. Low-resolution mass
spectrometry analyses were conducted on Waters Acquity UPLC and SQ systems. High-resolution mass
spectrometry analyses were conducted on an Agilent 6220 TOF mass spectrometer in positive
electrospray mode. The system was calibrated to greater than 1 ppm accuracy across the mass range prior
to analyses. The samples were separated using UHPLC on an Agilent 1200 system prior to mass
spectrometric analysis.
All water used in these preparations was purified water. Solvents used were commercial anhydrous grades
that were used as received. Residual water was determined using Karl Fischer titration (KF). Residual
solvents were determined by GCMS. Chiral purity determinations were made by gas chromatography.47
ACS Paragon Plus Environment

Page 21 of 46
Organic Process Research & Development
1
2
3
4
5
6
All previously reported compounds prepared in this work were identical in their compound
characterization to properties reported previously.2,7
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
(7aS)-3,3-dimethyl-6-(phenylsulfinyl)tetrahydro-3H,5H-pyrrolo[1,2-c]oxazol-5-one (12): A clean and
dry 8000 L glass-lined reactor was evacuated to 0.05 to 0.08 MPa and then filled with nitrogen to normal
pressure. This was repeated 3 times. The reactor was sampled for oxygen content to ensure it was ≤ 1.0%.
The reactor was rinsed with THF (83.0 kg), and the rinsing liquor was discharged to establish a residual
water content of  0.05%. Sodium hydride (60% in oil, 109.6 kg containing 65.8 kg NaH, 2742 mol, 2.0
equiv) was added at 15 to 30 C under nitrogen. The mixture was adjusted to 10 to 20 C, and THF (2698
kg) was added. After the addition, the mixture was bubbled with nitrogen for 1 h, then heated to 50 to 55
C and maintained at that temperature for 0.5 to 1 h.
A solution of 4 (205.9 kg, 1327 mol, 1.0 equiv) in THF (767 kg) was prepared in a dry 3000 L reactor
under nitrogen. The mixture was stirred until it was homogenous, then was added into the 8000 L reactor
at 50 at 55 C at a rate of 110 to 200 kg/h. Once the addition was complete, the mixture was maintained at
50 to 55 C for 0.5 to 1 h, then cooled to 45 to 50 C. A solution of methyl benzenesulfinate (319.6 kg,
2046 mol, 1.5 equiv) in THF (915.6 kg) was added into the mixture in five portions at a rate sufficient to
maintain the temperature at 45 to 55 C, while allowing 0.5 to 1 h between additions to allow hydrogen
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 46
1
2
3
4
5
6
7
8
9
evolution to subside between additions. Once the addition was complete, the mixture was stirred at 35 to
48 C with periodic nitrogen sparging. After about 6 h, the mixture was sampled every hour for HPLC
analysis until the content of 4 was ≤ 1%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Once the reaction was complete by HPLC analysis, the mixture was cooled to -5 to 0 C. While
maintaining the temperature at -5 to 5 C, water (414.4 kg) was added into the mixture to quench the
reaction. Then the mixture was stirred for 20 to 30 min at -5 to 5 C. The mixture was purged with
nitrogen until the hydrogen content was ≤0.5%. Acetic acid (227.4 kg, 3784 mol, 2.8 equiv) was added at
a rate of 20 to 50 kg/h to adjust the verified pH to 4 to 6. Stirring was then continued for 20 to 30 min at -
5 to 5 C. The mixture was then concentrated at ≤ 35 C under reduced pressure (≤ 0.08 MPa) until the
remaining volume was between 600 and 1000 L. Water (2065.0 kg) and CH2Cl2 (2755.4 kg) were added,
stirred for 0.5 to 1 h at 15 to 30 C, and settled for 0.5 to 1 h before separation of the CH2Cl2 phase. The
aqueous phase was extracted with additional CH2Cl2 (1374.7 kg) and stirred and settled as before. The
combined CH2Cl2 extracts were concentrated at ≤ 40 C under reduced pressure (≤ 0.08 MPa) until the
remaining volume was between 600 and 800 L. MTBE (1378.8 kg) was added and the mixture was stirred
concentrated as before until the remaining volume was between 600 and 1000 L. The mixture was
sampled for residual CH2Cl2 and KF analysis until residual CH2Cl2 was ≤ 10% and water was ≤ 0.15%.
ACS Paragon Plus Environment

Page 23 of 46
Organic Process Research & Development
1
2
3
4
Heptane (847.0 kg) was added and the mixture was stirred for 10 to 20 min before cooling to 5 to 10 C
5
6
7
8
9
for 1 to 2 h. The mixture was filtered with a stainless steel centrifuge. The filter cake was rinsed with
MTBE (307.2 kg) until purity analysis by HPLC showed ≥ 95% purity. The filter cake was dried in a
rotary conical dryer at 35 to 45 C under reduced pressure (≤ 0.08 MPa) until sampling for residual
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 and KF analysis showed residual MTBE ≤ 0.5%, residual heptane ≤ 0.5%, and residual water ≤
0.1%. There was obtained 268.1 kg (72%) of 12 as a fine white powder, melting point 153 - 155 C. Both
the 1H and 13C NMR data were consistent with the presence of two diastereomers in present in equal
proportions. 1H NMR (500 MHz, CD3OD) δ 7.68 - 7.73 (m, 2 H), 7.53 - 7.65 (m, 8 H), 4.39 (dd, J =
10.76, 8.31 Hz, 1 H), 4.16 - 4.28 (m, 2 H), 4.08 (dd, J = 8.31, 5.62 Hz, 1 H), 3.51 (dd, J = 9.29, 8.31 Hz, 1
H), 3.39 (dd, J = 9.78, 8.31 Hz, 1 H), 2.40 (ddd, J = 14.43, 7.34, 2.20 Hz, 1 H), 2.23 (ddd, J = 12.84,
10.64, 8.07 Hz, 1 H), 2.01 (ddd, J = 14.43, 10.03, 7.34 Hz, 1 H), 1.69 (ddd, J = 12.84, 8.19, 6.60 Hz, 1
H), 1.65 (s, 3 H), 1.64 (s, 3 H), 1.45 (s, 3 H), 1.43 (s, 3 H). 13C{1H} NMR (126 MHz, CD3OD) δ =
167.6, 166.9, 142.3, 141.9, 133.0, 132.6, 130.8, 130.7, 125.6, 125.3, 93.4, 93.4, 73.9, 73.0, 71.1,
70.7, 62.6, 60.3, 27.3, 27.0, 24.2, 23.9, 19.3, 19.2. IR (neat): 1690, 1043 cm-1. Anal. Calcd for
C14H17NO3S: C, 60.19; H, 6.13; N, 5.01; S, 11.48. Found: C, 60.12; H, 6.25; N, 5.02; S, 11.67. HRMS
(ESI/QTOF) m/z: [M+H+] Calcd for C14H17NO3S 280.1002; Found: 280.1000.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 46
1
2
3
4
5
6
7
8
9
(S)-3,3-dimethyl-1,7a-dihydro-3H,5H-pyrrolo[1,2-c]oxazol-5-one (6): A clean and dry 8000 L glass-
lined reactor was prepared under N2 as described above. The reactor was charged with toluene (379.9 kg)
and sampled to establish residual water at ≤ 0.05%. Additional toluene (2145.9 kg) was added to the
reactor and stirred for about 10 min. Anhydrous Na2HPO4 (447.2 kg, 3149 mol, 3.0 equiv) was added at
15 to 30 C, followed by 12 (290.5 kg, 1064 mol, 1.0 equiv). The mixture was heated to 80 to 85 C, and
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
maintained for 1 h, then heated further to 95 to 115 C. After 14 h, HPLC analysis showed ≤ 1.0% of 12
remaining. The mixture was cooled to 10 to 30 C and filtered through a stainless steel filter, and the
reactor was rinsed with toluene (507.3 kg). The filter cake was washed with the rinsing liquor and the
combined filtrates were concentrated at ≤ 45 C under reduced pressure until the remaining volume was
between 600 and 1200 L. The solution was cooled to 15 to 25 C, diluted with heptane (3979.6 kg),
stirred for 30 min, and cooled to -5 to 5 C. The mixture was extracted three times with equal portions of
a solution prepared from methanol (663.4 kg) and water (4834.0 kg), pre-cooled to -5 to 5 C. Each
extraction operation was stirred for 30 to 40 min and then allowed to settle for 15 to 30 min before
separation of the phases. The combined aqueous phases were extracted with CH2Cl2 (3857.0 kg) for 20 to
30 min, then NaCl (86.8 kg) was added and stirred for 5 to 10 min to facilitate phase separation. The
aqueous phase was extracted twice more with CH2Cl2 (1937.8 kg, 1886.6 kg). Each time the mixture was
stirred for 15 to 30 min and allowed to settle for 15 to 30 min prior to phase separation. The combined
ACS Paragon Plus Environment

Page 25 of 46
Organic Process Research & Development
1
2
3
4
CH2Cl2 extracts were adjusted to 10 to 20 C and dried with anhydrous Na2SO4 (100.0 kg) with stirring
5
6
7
8
9
for 30 min until KF analysis showed ≤ 1.0% water remaining. The mixture was filtered with a stainless
steel filter and the filter was washed with CH2Cl2 (771.4 kg). The combined CH2Cl2 filtrates were
concentrated at ≤ 35 C under reduced pressure (≤ 0.06 MPa) until the remaining volume was between
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
450 and 700 L. THF (777.0 kg) was added and the distillation was continued until the remaining volume
was between 450 and 700 L. Two additional cycles of THF addition (775.0 kg, 776.8 kg) and distillation
were conducted until analyses for residual water, MeOH and CH2Cl2 showed each to be present at ≤
0.10%. The resulting solution of 6 was cooled to 15 to 25 C, the reactor was rinsed with additional THF
(101.4 kg), the rinsing liquor was combined, and the solution was chilled to -30 to -10 C.
(7R,7aS)-7-ethyl-3,3-dimethyltetrahydro-3H,5H-pyrrolo[1,2-c]oxazol-5-one (7): A clean and dry 8000 L
glass-lined reactor was prepared under N2 as described above. The reactor was charged with THF (174.0
kg) at 15 to 30 C and sampled to establish residual water ≤ 0.05%. Additional THF (532.6 kg) was added
and N2 was bubbled at bottom of reactor to degas for 20 min, after which CuI (8.0 kg, 73.1 mol, 0.07
equiv) was added through a solid addition funnel. After addition, the funnel was rinsed with THF (4.0
kg). The mixture was cooled to -65 to -60 C, after which 1 M ethyl magnesium bromide in THF (1329.0
kg, 1245 mol, 1.2 equiv) was added into the reactor at a rate of 400 to 800 kg/h while maintaining the
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 46
1
2
3
4
5
6
7
8
9
temperature at -45 to -65 °C. Additional THF (10.6 kg) was used to rinse the addition port and the
mixture was stirred for 40 to 60 min before being cooled to -78 to -65 C. Me3SiCl (124.0 kg, 1141 mol,
1.1 equiv) was added dropwise into the mixture. THF (10.0 kg) was used to rinse the addition line and the
mixture was stirred for 1 to 2 h at -78 to -65 C. The pre-cooled solution of 6 was added while
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
maintaining the temperature at -78 to -65 C. The vessel containing the solution of 6 was rinsed twice
with THF (47.2 kg, 47.6 kg) and the washes were added into the reaction mixture. After 4.5 h, GCMS
analysis showed that ≤ 1.0% of 6 remained. The reaction mixture was added into a pre-cooled (-10 C)
solution of NH4Cl (95.4 kg) in water (953.0 kg) at a rate of 50 to 200 kg/h. The reactor was washed with
THF (48.4 kg) and the wash liquor was combined with the mixture. The mixture was warmed to 10 to 20
C and stirred for 2 h before being filtered through Celite® (23.8 kg). The filter cake was washed with
MTBE (590.3 kg) and the filtrate was stirred for 1 h at 10 to 30 C, then settled for 1 h before separation.
The aqueous phase was further extracted with twice with MTBE (355.6 kg, 355.4 kg), stirring and settling
before separation as before. The combined organic phases were washed with a brine solution prepared
from NaCl (200.6 kg) and water (804.0 kg) prior to concentration at ≤ 40 C under reduced pressure (≤ -
0.08 MPa) until the remaining volume was 600 to 800 L. The mixture was filtered through Celite® (40.0
kg). The filter cake was washed five times with MTBE (236.4 kg, 235.5 kg, 234.9 kg, 235.0 kg, 234.4
kg). The last washing was sampled for 7 content and found to be < 0.1%. The combined organic phases
ACS Paragon Plus Environment

Page 27 of 46
Organic Process Research & Development
1
2
3
4
were concentrated at ≤ 40 C under reduced pressure (≤ 0.08 MPa) until the remaining volume was 350 to
5
6
7
8
9
500 L. THF (694.8 kg) was added and concentration was repeated until the remaining volume was 350 to
500 L. Further addition of THF (695.0 kg, 694.8 kg) and concentration were conducted until analyses
showed residual water ≤ 0.05% and MTBE ≤ 0.5%. Repeated additions of heptane (329.1 kg, 164.4 kg,
164.0 kg, 163.8 kg, 165.2 kg) and concentration to the same residual volume were conducted, after which
THF (401.0 kg) was added and the mixture was cooled to 15 to 25 C. There was obtained 635.0 kg of
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 solution of 7 containing 165.1 kg (87%) of 7 of 94% HPLC purity and 100% ee.
(6R,7S,7aS)-7-ethyl-6-hydroxy-3,3-dimethyltetrahydro-3H,5H-pyrrolo[1,2-c]oxazol-5-one (14): Two
solutions of 7 prepared as above, containing in total 318.2 kg (1736 mol) of 7, were converted to 14 in six
approximately 53 kg batches as previously described.36 Following workup and purification there was
obtained 241.1 kg (1210 mol, 70%) of 14 as a solution in toluene containing approximately 24% by
weight of 14.
(6S,7S,7aS)-7-ethyl-6-fluoro-3,3-dimethyltetrahydro-3H,5H-pyrrolo[1,2-c]oxazol-5-one (8): A clean
and dry 5000 L glass-lined reactor was prepared under N2 as described above. A solution of 14 (118.1 kg,
593 mol, 1.0 equiv) in toluene (372.6 kg) was added, followed by additional toluene (704.0 kg). After
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 46
1
2
3
4
5
6
7
8
9
stirring for 20 min, perfluorobutane sulfonyl fluoride (231.6 kg, 767 mol, 1.3 equiv) and TMEDA (95.6
kg, 823 mol, 1.4 equiv) were added at 20 to 30 C. After 23 h, HPLC analysis showed 7% of 14
remaining and the concentration of 14 did not change with further sampling. Water (1170.7 kg) was
added at 20 to 30 C, the mixture was stirred for 1 h and allowed to settle for 2 h before separation. 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
aqueous phase was extracted twice more with additional toluene (1011.9 kg, 509.4 kg), stirring and
settling as before. The combined toluene solutions were transferred into a clean 5000 L reactor and
washed twice with water (1189.9 kg, 354.8 kg), stirring and settling as before. The toluene solution was
concentrated at ≤ 50 C under reduced pressure (≤ 0.08MPa) until the remaining volume was between
150 and 250 L. Precipitated solids were dissolved by heating and stirring at 45 to 55 C prior to the
addition of heptane (237.6 kg). The mixture was sampled for toluene content until the remaining toluene
content was between 10 and 18%. Once the toluene content was within specification, the mixture was
cooled to 0 to 10 C and maintained at that temperature for 2 h. Additional heptane (321.9 kg) was
charged into the mixture over 5.5 h, then maintained at 0 to 10 C for 2 h. The mixture was then cooled to
-5 to -10 C over 2.25 h before being filtered with a centrifuge. The filter cake was washed with heptane
(79.8 kg) before being transferred into a rotary conical dryer. It was dried at 15 to 30 C, under reduced
pressure (≤ 0.06MPa) for 11 h until analyses for residual solvents detected no residual toluene or heptane.
There was obtained 86.5 kg (72%) of 8 of 98% HPLC purity as brown crystals. 1H NMR (400MHz,
ACS Paragon Plus Environment

Page 29 of 46
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
DMSO) δ = 5.40 (dd, J = 7.4, 51.5 Hz, 1H), 4.09 - 3.98 (m, 2H), 3.67 - 3.56 (m, 1H), 3.67 - 3.56 (m, 1H),
2.75 - 2.64 (m, 1H), 1.54 (s, 3H), 1.51 - 1.40 (m, 1H), 1.36 (s, 3H), 1.34 - 1.21 (m, 1H), 0.87 (t, J = 7.2
Hz, 3H). 13C{1H} NMR (101MHz, DMSO) δ = 164.9 (d, J = 24.2 Hz), 92.4 (d, J = 195.1 Hz), 90.3, 64.1,
57.3, 41.2 (d, J = 16.1 Hz), 26.1, 23.1, 15.0 (d, J = 8.8 Hz), 12.5 (d, J = 2.9 Hz). 19F NMR (H decoupled,
376 MHz, CDCl3)  -199.61.
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
(3S,4S,5S)-4-ethyl-3-fluoro-5-(hydroxymethyl)pyrrolidin-2-one (2): A clean and dry 1000 L glass-lined
reactor was prepared under N2 as described above and charged with CH3CN (349.4 kg). A solution of 8
(73.5 kg, 365 mol, 1.0 equiv) in CH3CN (126.4 kg) was added and stirred for 1 h at 10 to 30 C, after
which water (44.2 kg, 6.7 equiv) was added. After stirring for 30 min, CF3CO2H (8.3 kg, 73 mol, 0.2
equiv) was added and stirred for 30 min before being heated to reflux. After 5 h, HPLC analysis showed ≤
0.1% of 8 remaining. The mixture was cooled to 25 to 35 C and transferred to a clean reactor, along with
CH3CN (73.4 kg) used to wash the reactor. The CH3CN solution was combined with a second CH3CN
solution resulting from similar processing of 86.5 kg (430 mol) of 8 for work up and crystallization. The
combined solutions were concentrated at ≤ 40 C under reduced pressure (≤ 0.08MPa) until the remaining
volume was between 180 and 230 L. Three additional portions of CH3CN (380.7 kg, 371.8 kg, and 376.9
kg) were added and distilled similarly. KF analysis showed 0.3% residual water present. Two portions of
ACS Paragon Plus Environment