Development of the Enabling Route for Glecaprevir via Ring-Closing Metathesis

Article abstract of DOI:10.1021/acs.oprd.9b00469

Glecaprevir was identified as a potent HCV NS3/4A protease inhibitor, and an enabling synthesis was required to support the preclinical evaluation and subsequent Phase I clinical trials. The enabling route to glecaprevir was established through further development of the medicinal chemistry route. The key steps in the synthesis involved a ring-closing metathesis (RCM) reaction to form the 18-membered macrocycle and a challenging fluorination step to form a key amino acid. The enabling route was successfully used to produce 41 kg of glecaprevir, sufficient to support the preclinical evaluation and early clinical development.

Full text of DOI:10.1021/acs.oprd.9b00469

Subscriber access provided by AUT Library
Full Paper
Development of the Enabling Route for
Glecaprevir via Ring Closing Metathesis
Russell D Cink, Kirill Lukin, Richard D Bishop, Gang Zhao, Matthew J Pelc, Timothy
B Towne, Bradley D Gates, Matthew M Ravn, David R Hill, Chen Ding, Steven
C. Cullen, Jianzhang Mei, M. Robert Leanna, Jeremy Henle, Jose G Napolitano,
Nandkishor Nere, Shuang Chen, Ahmad Y. Sheikh, and Jeffrey M. Kallemeyn
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00469 • Publication Date (Web): 27 Dec 2019
Downloaded from pubs.acs.org on December 29, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Development of the Enabling Route for Glecaprevir
via Ring Closing Metathesis
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Russell D. Cink*, Kirill A. Lukin, Richard D. Bishop, Gang Zhao, Matthew J. Pelc, Timothy B.
†
Towne, Bradley D. Gates, Matthew M. Ravn, David R. Hill, Chen Ding, Steven C. Cullen,
Jianzhang Mei, M. Robert Leanna, Jeremy Henle, José G. Napolitano, Nandkishor K. Nere,
Shuang Chen, Ahmad Sheikh, and Jeffrey M. Kallemeyn
Process Research & Development, AbbVie Inc., 1401 Sheridan Road, North Chicago, IL, 60064
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
For Table of Contents Only (first is 1200 dpi, second is 300 dpi, third is chemdraw)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
F
F
F
F
N
N
N
N
Enabling
RCM
O
O
O
O
Route
O
O O
H
O
O
N
N
S
N
OMe
O
N
O
H
O
O
O
O
N
H
N
H
F
F
Glecaprevir
KEYWORDS
glecaprevir, ABT-493, ring closing metathesis, HCV
ACS Paragon Plus Environment

Page 3 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Abstract: Glecaprevir was identified as a potent HCV NS3/4A protease inhibitor and an
enabling synthesis was required to support the pre-clinical evaluation and subsequent Phase I
clinical trials. The enabling route to glecaprevir was established through further development of
the medicinal chemistry route. The key steps in the synthesis involved a ring closing metathesis
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(RCM) reaction to form the 18-membered macrocycle, and a challenging fluorination step to
form a key amino acid. The enabling route was successfully used to produce 41 kg of
glecaprevir, sufficient to support the pre-clinical evaluation and early clinical development.
Introduction
The hepatitis C virus (HCV) is a blood borne disease estimated to affect between 71 and 185
1
million people throughout the world. Persons infected with HCV can show no signs of the
disease for decades. Left untreated HCV can lead to liver failure, liver cancer, and ultimately
death. One of the challenges in developing treatments is the genetic heterogeneity of HCV, with
six major genotypes identified. Fortunately, the development of combinations of direct acting
antiviral agents in recent years has resulted in multiple treatment options that represent a curative
therapy for all major HCV genotypes.2
In December 2006, Abbott Laboratories, which subsequently split to form AbbVie, began a
collaborative effort with Enanta Pharmaceuticals focused on the development of HCV protease
inhibitors. The collaboration led to the development of ABT-450 as part of AbbVie’s first
3
generation products for HCV treatment. The first generation products were multi-drug
combinations that were limited to the treatment of patients with genotypes 1a, 1b, and 4. A
second generation program was initiated to discover and develop a treatment with broad
genotypic activity that required fewer active ingredients in the combination. Glecaprevir was
4
identified as a potent HCV NS3/4A protease inhibitor with pan-genotypic activity. Glecaprevir
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
was developed in combination with pibrentasvir, an NS5A inhibitor which also exhibits pan-
genotypic activity. The combination of glecaprevir and pibrentasvir was approved for the
treatment of chronic hepatitis C virus genotypes 1–6.5
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
An enabling synthesis of glecaprevir was required to support the pre-clinical evaluation and
subsequent Phase I clinical trials. The retrosynthetic analysis for the medicinal chemistry route to
glecaprevir (1), as shown in Scheme 1, began with an initial disconnection of the amide bond
dissecting glecaprevir (1) into amino sulfonamide 2 and macrocyclic acid 3. The amino
sulfonamide 2 was prepared from sulfonamide 7 and amino acid 8. The key step in the synthesis
of amino acid 8 was the fluorination of the precursor aldehyde 9. The key step in the synthesis of
macrocyclic acid 3 was a ring closing metathesis reaction (RCM) of diene 4, which was in turn
derived from coupling of carbamate 5 with amine 6.6
Scheme 1. Retrosynthetic Analysis of the Enabling Route to Glecaprevir
F
F
F
F
F
F
N
N
N
N
N
N
O
O
O
O
O
O
O
O O
S
O
O O
S
H
O
H2N
O
N
N
N
H
N
OH
O
N
OMe
O
N
+
O
O
H
O
O
O
O
O
O
N
H
F
N
H
F
N
H
F
F
2
4
3
Glecaprevir, 1
F
F
N
O
O
O
O
O
O
BocHN
Boc2N
OEt
O
O
N
OH
S
+
O
OH
+
H2N
HN
H
F
HN
OMe
O
7
F
O
8
9
5
6
Given the structural complexity of glecaprevir (1) and the time constraints of the project, the
strategy for the preparation of initial batches of glecaprevir (1) was to utilize the same major
ACS Paragon Plus Environment

Page 5 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
disconnections as the existing medicinal chemistry route while implementing improvements to
enhance the scalability of the synthesis. An evaluation of the medicinal chemistry route revealed
two major challenges and two minor challenges for the large-scale synthesis of 1.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The first major challenge was the fluorination of aldehyde 9 which proceeded in less than 50%
yield, employed the hazardous reagent diethylaminosulfur trifluoride (DAST), and resulted in
7
recovery of unreacted starting material upon chromatographic purification. The first key
objective was the development of a fluorination procedure suitable to meet the short-term
material requirements of the program. Due to the hazardous nature of DAST and similar
reagents, this fluorination step was recognized as a serious limitation for this synthesis of amino
8
acid 8. The strategy was to utilize this route for the initial production of 8 while an improved
synthesis was being developed for larger scale production. The development of the synthetic
route used for large scale production will be detailed in a subsequent publication.9
The second major synthetic challenge was the RCM reaction used to form macrocyclic acid 3
which proceeded in 59% yield, employed 9 mol % of the Zhan 1B catalyst, and required
1
0
chromatographic purification. Based on prior experience with similar substrates, a challenging
purification was expected due to the likely formation of dimeric impurities and the cis alkene
isomer. Therefore, the second key objective was the development of a reliable procedure for the
RCM reaction that minimized the formation of dimeric impurities and the cis alkene isomer.
The two minor challenges were the reported low overall yields for the synthesis of amine 6
(34%) and carbamate 5 (12%).10 The third key objective was therefore to develop robust
processes that increased the yield for both amine 6 and carbamate 5. With these three objectives
in mind, the development of the enabling route to glecaprevir (1) began.
Results and Discussion
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Synthesis of the Amino Acid 8
As shown in Scheme 2, the synthesis of aldehyde 9 began with the Boc amino ester 10,
1
1
synthesized in 6 steps and 29% yield from glycine ethyl ester. The Boc amino ester 10 was first
protected with an additional Boc group and the alkene was then subjected to osmium tetroxide
catalyzed cleavage with sodium periodate to furnish aldehyde 9 in 75% overall yield.
Scheme 2. Synthesis of Aldehyde 9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
O
BocHN
1) Boc2O, DMAP, THF
Boc2N
OEt
OEt
2
) NaIO4, OsO4O
,4-Dioxane
2,6-Lutidine
1
H
O
7
5% from 10
10
9
The subsequent fluorination reaction to form difluoro ester 11 proved to be particularly
challenging. As shown in Table 1, the formation of a significant amount of side product 12 along
with a small amount of alkene impurity 13 was observed. The formation of 13 is presumably the
result of cyclopropane ring fragmentation of the initially formed mono-fluoro intermediate with
concomitant addition of fluoride α to the ester. The formation of 12 is presumably the result of
the ester carbonyl oxygen trapping the initially formed mono-fluoro intermediate, followed by
1
2
fluoride addition to the ester carbonyl. Side product 12 was stable under the reaction conditions
but was not stable to reverse phase HPLC or silica gel chromatography due to hydrolysis back to
the starting aldehyde 9. The formation of side product 12 was the reason for recovery of
unreacted starting material 9 upon chromatographic purification rather than incomplete
1
conversion in the reaction. As a result, the reactions were best monitored by H NMR to assess
1
3
the conversion of aldehyde 9, along with the amounts of 11, 12 and 13 formed. A parameter
screen of the fluorination reaction was conducted in an effort to maximize the yield of 11 by
minimizing the formation of side products 12 and 13, and the results are summarized in Table 1.
ACS Paragon Plus Environment

Page 7 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Table 1. Screen of Fluorination Parameters for Aldehyde 9
O
O
F
O
OEt
O
Boc2N
Boc2N
fluorination
conditions
Boc2N
Boc2N
OEt
F
OEt
OEt
H
20 °C
H
F
F
Ha
O
F
F
H
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
9
11
12
13
1
Ratio by H NMR
Entry
Reagent
Solvent
2,6-Lutidine
1
1
12
13
1
2
3
4
5
6
7
8
9a
Methyl DAST (4 equiv)
Deoxo-Fluor® (4 equiv)
Deoxo-Fluor® (2 equiv)
DAST (4 equiv)
CDCl3
CDCl3
CDCl3
CDCl3
CDCl3
CDCl3
MeCN
Toluene
Heptane
CH Cl
0.25 equiv
none
50
45
31
50
51
50
15
50
39
57
55
53
46
52
69
41
45
47
56
46
52
39
41
40
4
3
0.25 equiv
none
0
9
DAST (4 equiv)
0.1 equiv
0.2 equiv
none
4
DAST (4 equiv)
3
DAST (4 equiv)
29
4
DAST (4 equiv)
none
DAST (4 equiv)
none
2
1
0a
DAST (4 equiv)
0.25 equiv
0.2 equiv
0.25 equiv
3
2
2
2
2
1
1
DAST (3.2 equiv)
DAST (2.5 equiv)
CH Cl
4
2
1
2a
CH Cl
4
2
(a) A small amount starting aldehyde 9 was present and accounts for the balance for these experiments.
SF3
MeO
N
SF3
N
SF3
N
SF3
N
SF2
BF4
O
N
SF2
BF4
MeO
Deoxo-fluor®
t-Bu
Methyl
DAST
DAST
Fluolead™
XtalFluor-E®
XtalFluor-M®
A screen of fluorination reagents was conducted and DAST, methyl DAST, and Deoxo-Fluor®
gave the best conversion of aldehyde 9 to the difluoro ester 11. Negligible conversion was
observed with the alternative fluorination reagents Fluolead™, XtalFluor-E®, and XtalFluor-
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
M®.14 DAST and Methyl DAST gave essentially the same profile (entries 1 and 6), whereas
Deoxo-Fluor® showed less of product 11 and more of side product 12 (entries 2 and 3). The
addition of a sub-stoichiometric amount of 2,6-lutidine was observed to reduce the level of
alkene impurity 13 formed in the reaction for both DAST and Deoxo-Fluor® (entry 2 compared
to 3, and entry 4 compared to 5 and 6). The addition of 2,6-lutidine also significantly slowed the
reaction rate. Both observations are presumably due to the neutralization of HF in the reaction
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
mixture. Other additives tested included EtOH, water, HF pyridine, BF etherate, silica gel, and
3
tetrabutylammonium difluorotriphenylsilicate (TBAT); all showed no improvement in the
profile. Additionally, attempts to convert the side product 12 to the desired product 11 were
unsuccessful when higher temperatures or more forcing conditions, including neat DAST, were
employed. A solvent screen (entries 7 to 10) showed that toluene was comparable to CDCl , but
3
CH Cl produced the highest level of the desired product 11. Evaluation of the reagent
2
2
stoichiometry for the fluorination reaction (entries 10-12) showed the amount of DAST could be
reduced to without impacting the product profile significantly. Based on these results, CH Cl
2
2
was selected as the reaction solvent and DAST (2.5 equiv) was selected as the fluorination
reagent.
The primary challenge moving forward was the workup and in particular how to address the
high level of the side product 12 formed in the reaction. It was found that upon hydrolysis of side
product 12 back to starting material, the aldehyde 9 could be separated and recovered through
conversion to the bisulfite adduct, as outlined in Scheme 3. For the process workup upon scale-
up, the reaction mixture was quenched into aqueous K HPO to consume the remaining DAST
2
4
and extracted with heptane to give the crude product mixture containing 11, 12 and 13. The
solution was then concentrated and mixed with aqueous KH PO to hydrolyze side product 12
2
4
ACS Paragon Plus Environment

Page 9 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
back to the starting aldehyde 9. The aldehyde 9 was then converted into the bisulfite adduct and
separated into the aqueous layer while the product 11 and impurity 13 were extracted into
heptane. The bisulfite adduct in the aqueous layer was then treated with sodium carbonate to
regenerate the aldehyde 9 which crystallized from the mixture in 35% recovery relative to the
starting amount of 9. The average yield of 11 in the fluorination step was 40% over numerous
batches. It was demonstrated that the recovered aldehyde 9 could be recycled through the
fluorination step in the same yield.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Synthesis of Amino Acid 8
O
F
O
O
1
) DAST, 2,6-Lutidine
CH Cl , 20 °C
OEt
O
Boc2N
Boc2N
Boc2N
BocHN
2
2
OEt
OEt
OH
LiOH
2) aq. K2HPO4
heptanes
+
EtOH
THF
H
F
F
F
H
O
F
F
50 °C
9
12
11
8
3
4
) aq. KH2PO4
) NaHSO3
40%
yield from 9
80%
5) Na2CO3
Separated as bisulfite adduct
35% recovery of aldehyde 9
The difluoro ester 11 was then converted to the amino acid 8 in 80% yield. The use of LiOH
was found to be important in cleanly removing the Boc group as NaOH resulted in a 2:1 mixture
of mono-Boc and di-Boc acids at 20 °C with higher temperature resulting in lower yields.
Impurity 13 underwent decomposition in the saponification and the resulting daughter impurities
were rejected in the crystallization of amino acid 8.
The yield of amino acid 8 was 24% from the Boc amino ester 10, not accounting for a yield
increase from recycle of recovered aldehyde 9 through the fluorination step. Although the
process required a complicated workup sequence for the fluorination of aldehyde 9,
chromatography was eliminated from the process, with purification being accomplished by
crystallization of aldehyde 9 and amino acid 8. Taking into consideration the reported 29% yield
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
11
for the synthesis of Boc amino ester 10, the overall yield of amino acid 8 was 7%. The process
outlined in Scheme 2 and Scheme 3 was conducted for the production of multiple batches amino
acid 8. This process met the objective of providing sufficient material for the early development
of glecaprevir (1), but it was clear that this synthesis of amino acid 8 was not suitable for larger
scale production. As mentioned earlier, the development of the synthetic route used for large
scale production will be detailed in a subsequent publication.9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of Amino Sulfonamide 2
With the synthesis of amino acid 8 in hand, the conversion to amino sulfonamide 2 was
evaluated. As shown in Scheme 4, the coupling of amino acid 8 with the commercially available
sulfonamide 7 was accomplished using HATU and DMAP. The coupling reaction required
excess DMAP (4 equiv) to achieve good conversion and multiple aqueous washes to remove the
DMAP. The Boc deprotection proceeded smoothly with HCl and the yield of amino sulfonamide
2 was 89% yield for the two-step sequence.
Scheme 4. Synthesis of Amino Sulfonamide 2
O
1) HATU, DMAP
,6-Lutidine
MeCN
O
O
O
2
BocHN
O
O
H2N
HCl
S
OH
N
H
S
H2N
+
2) HCl in IPA
F
F
i-PrOAc
F
7
F
8
8
9% from 8
2
Synthesis of Carbamate 5
As shown in Scheme 1, the other main challenge was the synthesis of macrocyclic acid 3
which required the preparation of carbamate 5 and amine 6. The retrosynthesis of carbamate 5,
as shown in Scheme 5, began with the dissection into tert-leucine 15 and the cyclopentane
alcohol 14, linked through the carbamate functionality. The coupling of the two fragments was
planned through the use of a phosgene equivalent. The conversion of alcohol 14 into the
ACS Paragon Plus Environment

Page 11 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
corresponding chloroformate was the preferred approach as this method had been successfully
applied to the formation of carbamate linkages in structurally related targets. Conversion of tert-
leucine 15 into the corresponding isocyanate was also considered, but typically requires the use
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
of an excess of the alcohol (2 equiv) to achieve a good yield of the carbamate. In this case the
chiral alcohol 14 was the higher value material in the coupling, so it was preferable to use an
excess of tert-leucine 15. Based on prior experience with other related compounds, only a
modest excess (1.2 equiv) of tert-leucine 15 would be required to form the carbamate bond.
Scheme 5. Retrosynthesis of Carbamate 5
(
R)
O
CO H
2
O
H N
2
O
(R)
+
O
CO H
OH
2
HN
14
15
5
Evaluation of the medicinal chemistry route to intermediate 5 revealed some potential
improvements to enhance the scalability of the synthesis. The chiral cyclopentane diol fragment
1
0
was synthesized by an enzymatic resolution of the diacetate 16, as shown in Scheme 6. The
resolution of diacetate 16 produced the desired mono-acetate 18 in 35% yield and was
subsequently converted to the alcohol 14 in 40% yield. One of the known challenges with the
resolution of diacetates such as 16 is that the product undergoes further conversion to the diol; in
1
5
this case mono-acetate 18 is converted to the diol 19. This results in a lowering of the yield of
18, and requires careful control of the reaction conditions to minimize over-reaction to form 19
as well as purification by chromatography.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 6. Initial Synthesis of Alcohol 14
Amano
OAc Lipase
(S)
(R)
OAc
OR
(S)
(R)
buffer
OAc
±)
OAc
OH
(
35%
18 R = Ac
19 R = H
17
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
16
(
R)
(R)
OAc 1) NaH, allyl bromide
2) LiOH
OH
(R)
(R)
OH
O
40%
18
14
An alternate approach considered for the cyclopentane diol motif was to utilize a catalytic
enantioselective epoxide opening of cyclopentene oxide, with three options outlined in Scheme
7. This approach has the advantage of potential higher yields as the meso epoxide can
theoretically be quantitatively converted to a single stereoisomer. The cobalt salen catalyst
developed by Jacobsen was known to produce the benzoate 20, but in modest yield and chiral
purity.16 The oligomeric cobalt salen catalyst was reported to improve the yield and
enantioselectivity for similar substrates, but was not commercially available.17 The gallium
BINOL catalysts developed by Shibasaki were known to produce the aryl ether 21 in good yield
and chiral purity, but the high catalyst loading and additional deprotection steps were considered
1
8
problematic for expeditious scale up of the process. Zhao and co-workers reported an epoxide
hydrolase (BD10721) was capable of producing diol 19 in good chiral purity, but this approach
was not evaluated due to time constraints.19
Scheme 7. Alternate Syntheses of the Cyclopentyl Motif
(R)
OR
(R)
O
OH
Option 1: Co Salen catalyst, BzOH 20: R = Bz
Option 2: Ga BINOL catalyst, ArOH 21: R = Ar
Option 3: Epoxide Hydrolase
19: R = H
ACS Paragon Plus Environment

Page 13 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
In order to simplify the resolution and separation of the enantiomers, an alternate sequence was
developed as outlined in Scheme 8. The trans cyclopentane diol motif could be derived from
cyclopentene oxide, and therefore the synthesis was modified to perform the ring opening of
cyclopentene oxide with allyl alcohol to produce racemic, trans substituted alcohol 22. Although
some dimers and trimers were formed upon further reaction of alcohol 22 with cyclopentene
oxide, the product was easily purified by distillation. Using this approach, the two alcohol
functionalities of the cyclopentane were differentiated in preparation for an enzymatic resolution.
After conversion to acetate 23, the proposed enzymatic resolution (Table 2) would yield only
two products, the undesired (S,S) acetate 24, and the desired (R,R) alcohol 14.20
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 8. Synthesis of Acetate 23
Ac2O
DMAP
HO
O
O
O
BF3OEt2
Toluene
DIPEA
Heptane
OH
(±)
OAc
(±)
69%
22
23
A screen of nineteen enzymes was conducted, and the results are summarized in Table 2 for
the seven enzymes which showed conversion. The chiral purity of the product alcohol 14 was
initially determined by derivatization as the Mosher ester and analysis by GC-MS, with the
configuration determined by NMR analysis. The Codexis enzyme NZL-102-LYO, and the
immobilized form NZL-102-IMB, showed excellent conversion and chiral purity (entries 2 and
5
). The Fluka lipase from Mucor miehei (entry 7) also showed excellent chiral purity, but the
reaction rate was significantly slower than entries 2 and 5. The enzyme in entries 2 and 5 was the
same, simply either in lyophilized or immobilized form, and was Candida Antarctica Lipase B
2
1
used to manufacture the commercially available immobilized enzyme Novozym 435. As such,
Novozym 435 was selected for further development of the resolution of 23.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 2. Enzyme Screen for Resolution of Acetate 23
enzyme
(S)
(R)
O
O
O
aqueous
buffer
(S)
OAc
(R)
OAc
OH
(±)
24
14
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
23
Entry
Enzymea
ee Config. 14
1
2
3
4
5
6
7
Codexis NZL-101-LYO
Codexis NZL-102-LYO
Codexis NZL-107-LYO
Codexis NZL-101-IMB
Codexis NZL-102-IMB
Sigma esterase E3019
0
> 95%
(R,R)
0
0
> 95%
0
(R,R)
(R,R)
Fluka lipase from Mucor miehei > 95%
(a) Reactions conducted using 0.25 mmol of 23 (neat) with 2.5 mL of 100 mM pH 7 phosphate buffer at 35 °C.
After development of a chiral HPLC method to allow more accurate characterization of the
product, the resolution with Novozym 435 was found to be exceptionally robust, producing
alcohol 14 in 47% overall yield from 22 with ≥ 99.8% ee (Scheme 9). It is noteworthy that the
resolution showed no erosion of chiral purity of 14 even upon extended reaction times, indicating
essentially no conversion of acetate 24. As shown in Scheme 9, the mixture of acetate 24 and
alcohol 14 was carried forward, with 14 being converted to the chloroformate 25 upon reaction
with triphosgene. After an aqueous workup, the chloroformate 25 was isolated as a toluene
solution that also contained the unreacted acetate 24. The chloroformate 25 was coupled to tert-
leucine 15 under biphasic conditions to produce carbamate 5 which was separated in the aqueous
layer as the sodium carboxylate salt, with the unreacted acetate 24 being easily removed in the
toluene layer. This simple strategy facilitated the separation of the enantiomers produced in the
ACS Paragon Plus Environment

Page 15 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
enzymatic resolution without additional purification steps. The carbamate 5 was then isolated as
the dicyclohexylamine (NH(Cy) ) salt for purification.
2
Scheme 9. Conversion of Acetate 23 to Carbamate 5
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CO2H
(R)
(
R)
triphosgene,
2,6-lutidine
1) NaOH, H2N
water
O
O
O
Novozym 435
O
1
4
(R)
15
(R)
O
O
aqueous
KH2PO4
K HPO
OH
toluene
O
CO2H NH(Cy)2
OAc
±)
O
2
) i-PrOAc, Heptanes
aqueous workup
HN
(
Cl
25
2
4
(
S)
O
23
47% from 22
99.8% ee
3) NH(Cy)2
24
(S)
>
OAc
5
80% yield
from 14
The overall 5 step sequence produced carbamate 5 in 26% yield from cyclopentene oxide,
without chromatography and with only one distillation and one crystallization required for
purification. By switching to a resolution of acetate 23 instead of resolution of diacetate 16, the
enabling process met the objective of increasing yield of carbamate 5 by more than doubling the
10
1
2% yield of the initial route. Overall, the enabling process to carbamate 5 was simple, robust,
and suitable for large scale production.
Synthesis of Amine 6
The strategy for the synthesis of amine 6, as shown in Scheme 10, was to couple the
commercially available hydroxyproline 26 with chloro quinoxaline 27 through a S Ar reaction.
N
This approach had been successful on several structurally related targets. The synthesis of chloro
quinoxaline 27 was planned through condensation of diaminobenzene 29 with the keto ester 28
which contained the α-difluoro alkene functionality.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 10. Retrosynthesis of Amine 6
F
F
N
N
OH
F
F
O
N
N
+
N
OH
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cl
Boc
HN
OMe
O
27
O
26
6
F
F
O
O
H2N
2
+
EtO
28
H N
29
The synthesis of the amine 6, shown in Scheme 11, began with the commercially available
bromide 30 and ethyl glyoxylate 31. Bromide 30 proved to be the most straightforward source of
the required α-difluoro alkene functionality in keto ester 28. Lithium halogen exchange of 30 had
been reported but required cryogenic conditions, and the zinc species was shown to provide
2
2
moderate yields upon addition to aldehydes and ketones. The bromide 30 was known to form
2
3
an allyl indium species suitable for alkylation of aldehydes. The formation of the allyl indium
species of 30 in a mixture of THF and water proved to be the ideal conditions for the addition to
ethyl glyoxylate 31.24
ACS Paragon Plus Environment

Page 17 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Scheme 11. Synthesis of Amine 6
NH2
NH2
F
F
F
F
F
F
T3P
DMSO
H
O
O
O
N
F
F
OH
Indium
29
+
Br
OEt THF/water
31
84%
EtOAc
to 20 °C
N
O
OEt
O
OEt
Et N, AcOH
HO
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
0
30
32
86% from 32
33
28
OH
F
F
F F
N
N
N
N
F
F
SOCl2
DMF
N
N
N
OH
O
O
Boc
HCl, MeOH
33
O
i-PrOAc
26
Cl
i-PrOAc
60 °C
N
OH
HN
HCl
OMe
NaOtBu, DMF, 0 °C
2
7
96%
Boc
77% from 27
O
O
3
4
6
The formation of the allyl indium species was found to be sensitive to the mixing efficiency. In a
lab scale cylindrical reactor with overhead stirring using a retreat curve agitator blade, the
indium was observed to settle on the bottom of the reactor due to the high density of indium (7
g/mL), resulting in a 10% lower yield. In order to avoid this issue in larger scale batches, the
reaction was run in a reactor with an agitator nearly touching the bottom of the reactor.
Additionally, the reaction showed a 20% lower yield when exposed to air, so the reaction
solvents were degassed to exclude oxygen from the reaction. With these modifications the yield
of the alkylation of ethyl glyoxylate 31 was 84% based on the bromide 30 (93% based on 31). A
slight undercharge of ethyl glyoxylate (0.9 equiv) was used to minimize the formation of
daughter impurities in the downstream process. The oxidation of 32 to the ketone 28 was initially
conducted using either Dess-Martin periodinane or TPAP/NMO. Further evaluation revealed that
2
5
the oxidation of 32 proceeded cleanly using T3P to activate DMSO at 0 °C. In our hands, this
reagent combination was found to be a generally mild oxidation reaction suitable for use with
sensitive substrates and did not require added base to complete the oxidation. The switch to
T3P/DMSO also avoided the need for controls for residual ruthenium with TPAP, and the safety
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
concerns associated with Dess-Martin periodinane. The ketone 28 was then condensed in situ
with diaminobenzene 29 to give quinoxaline 33 which was isolated in 86% yield by
crystallization from the reaction mixture. The enabling synthesis of quinoxaline 33 shown in
Scheme 11 resulted in a 72% overall yield, and represents a significant increase compared to the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
0
4
1% yield reported for the initial procedure. The yield increase was the result of the improved
conditions for the indium coupling to form alcohol 32 and implementing the T3P/DMSO
oxidation to form ketone 28.
The quinoxaline 33 was subsequently converted to the chloro quinoxaline 27. Crystallization
of 27 offered minimal purification so instead a silica plug filtration was used to remove color and
polar impurities. The S Ar reaction of chloro quinoxaline 27 with hydroxyproline 26 proceeded
N
smoothly to form the acid 34 in nearly quantitative yield. The removal of the Boc protecting
group and conversion to the methyl ester was accomplished using HCl in methanol, allowing for
isolation of 6 as the hydrochloride salt from the reaction mixture after a solvent switch to
isopropyl acetate. The crystallization of 6 was straightforward on lab scale but proved more
challenging at large scale. The crystallization required removal of the excess HCl in order to
achieve supersaturation, and this was accomplished by chase distillation with methanol. The
efficiency of the HCl removal by chase distillation was reduced at large scale and required larger
amounts of methanol. The starting 7.5 equivalents of HCl in the reaction were reduced to
approximately 3 equivalents by chase distillation with methanol which induced crystallization. In
the end the amine 6 was isolated in an average 77% yield from chloro quinoxaline 27.
The amine 6 was produced in 53% overall yield from bromide 30, with purification by
crystallization at intermediates 33 and 6 and a silica plug filtration of intermediate 27. The yield
of amine 6 was increased by more than 50% compared to the 34% yield reported for the initial
ACS Paragon Plus Environment

Page 19 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
procedure. The enabling process met the objective of improving the yield for the synthesis of
amine 6 due to the increased yields of the indium coupling to form 32 and the T3P/DMSO
oxidation to form 28.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of the Macrocyclic Acid 3
With the efficient syntheses of 5 and 6 in hand, the preparation of macrocycle 35 could be
achieved by an amide coupling followed by a ring closing metathesis (RCM) reaction, as shown
in Scheme 12. The coupling of acid 5 with the amine 6 using HATU as the coupling reagent to
produce diene 4 proceeded in quantitative yield; a significant increase compared to the initially
1
0
reported 59%. The diene 4 was purified by a silica plug filtration to ensure removal of potential
catalyst poisons in preparation for the subsequent RCM reaction.
Scheme 12. Synthesis of the Macrocycle Ester 35
F
F
F
F
F
F
N
N
N
N
N
N
O
O
O
O
O
O
O
O
+
HATU
RCM
O
OH NHCy2
O
O
HN
N
OMe
N
OMe
DIPEA
MeCN
HN
OMe
O
O
O
O
O
HCl
O
O
100%
N
H
N
H
5
6
4
35
The RCM reaction proved to be a significantly challenging step in the synthesis. The reaction
required a high catalyst loading in order to achieve near complete conversion and produced
significant amounts of the cis isomer impurity 36 and the dimer impurity 37, along with a low
level of the Zhan 1B catalyst-derived impurity 38. The structures of the RCM reaction impurities
are shown in Figure 1.26 It should be noted that the dimer impurity 37 was formed at higher
levels early in the course of the reaction but gradually converted to product resulting in a lower
final level.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
SO2NMe2
F
F
N
N
O
F
N
N
F F
F
O
O
O
N
N
O
O
O
O
HN
O
O
O
O
O
O
N
OMe
N
OMe
NH
N
O
O
OMe
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
OMe
O
O
O
N
H
N
N
O
O
O
O
O
37
N
H
36
F
F
38
Figure 1. Impurities Formed in the RCM Reaction
A catalyst screen was conducted, shown in Table 3, and revealed that the Zhan 1B catalyst and
the Grubbs second generation catalyst performed the best (entries 1 and 2), giving the highest
conversion to 35 and similar levels of impurities 36 and 37. The other catalysts screened (entries
4
-8) showed either low yields of 35, higher levels of the cis isomer impurity 36 relative to 35, or
2
7
high levels of dimer impurity 37. Based on literature precedent, the addition of acetic acid was
also evaluated as a potential strategy to reduce the level of the cis isomer impurity 36 but this
2
8
offered no improvement (entries 3, and 6, 7, and 8). The Zhan 1B catalyst was selected for
further evaluation as it provided the highest level of product and the lowest level of starting
material.
ACS Paragon Plus Environment

Page 21 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Table 3. Catalyst Screen for the RCM Reaction
F
F
F
F
N
N
N
N
O
O
O
O
catalyst
toluene
36 = cis isomer
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
37 = dimer
O
O
N
OMe
N
OMe
O
O
40 °C
O
O
O
O
N
H
N
H
4
35
HPLC Area %
Catalyst AcOH
mol %) (equiv)
Ratio
35/36
Entry
Catalysta
(
4
35
36
37
1
2
3
4
5
6
7
Zhan 1B
10
10
5
0
0
7.9
7.5
7.6
-
0.4
4.6
82
79
75.2
-
10.4
10.5
9.9
-
3.1
3.1
8.7
-
nd
Grubbs 2 Generation
Zhan 1B
0.5
0
6.2
Zhan 1B resin supported
Grubbs C801
3
~100
62.6
24
6
0
-
-
-
37.4
25.1
7
Omega CS1
5
0.5
0.8
5.8
7.1
43.4
25
7.5
3.5
CatMETium® RF1
4
53
Hoveyda-Grubbs 1st
Generation
8
6
0.5
-
58.5
-
-
41.5
(a) Reactions conducted at 42 ± 7 mg of 4 in toluene (100 mL/g of 4) at 40 °C.
iPr
iPr
N
N
Mes
PCy3
Ru
Mes
N
N
Mes
N
N
N
N
Mes
Mes
Mes
Cl
Mes
Ph
N
N
Mes
Ph
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Ru
Ru
Ru
Ru
O
Cl
Cl
O
S
Cl
iPr
Ru
O
iPr
O
O
N
PCy3
PCy3
Cy3P
N
H
CF3
O
Grubbs
2nd Generation
Grubbs
C801
Hoveyda-Grubbs
1st Generation
Zhan 1B
Omega CS1
CatMETium® RF1
A solvent screen showed the best impurity profile was obtained with toluene as compared to
CH Cl , EtOAc, THF, and heptane/CH Cl . Through the screening experiments in Table 4 it was
2
2
2
2
determined that lowering the reaction temperature reduced the amount of cis isomer impurity 36
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
relative to the amount of product 35 (entries 1-4, ratio of 35/36). However, at temperatures below
4
0 °C, the reactions were slower, required more catalyst, and showed higher levels of starting
material 4 and the dimer impurity 37 (entries 3, 4). The reactions run below 40 °C frequently
stalled and required multiple catalyst charges and extended reaction times to achieve reasonable
conversion. As a result, the reaction temperature was set at 40 °C to minimize the cis isomer
impurity 36 while allowing for near complete consumption of the starting material in a
reasonable timeframe. At this temperature, the reaction was complete in approximately 20 h with
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
0% of the cis isomer 36 and 5% of the dimer impurity 37 (entry 7). A high catalyst loading of 8
mol % relative to 4 was required due to the reduced reactivity of the α-difluoro alkene in the
2
9
substrate. At lower catalyst loadings the reaction would stall out and require additional catalyst
charges to restart the reaction, ultimately reaching 8 mol % of catalyst.
ACS Paragon Plus Environment

Page 23 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Table 4. Parameter Screen for the RCM Reaction
F
F
F
F
N
N
N
N
O
O
O
O
Zhan 1B
toluene
36 = cis isomer
37 = dimer
38 = ligand impurity
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
O
+
N
OMe
N
OMe
O
O
O
O
O
O
N
H
N
H
4
35
HPLC Area %
35 36 37
Temp Toluene Reaction Addition Zhan 1B Ratio
Entry
(°C) (mL/g 4) Time (h) Time (h) (mol %) 35/36
4
38
1 b
2 b
3 b
4c
80
40
30
22
40
40
40
100
100
100
100
20
7a
24
100
82
45
50
26
4
6
7
-
8
10
13
10
8
4.8
8.3
9.0
1.5 74.9 15.5 6.1 1.5
2.2 81.7 9.8 5.1 0.8
8.5 73.5 8.2 6.4 1.9
11.1 10.7 74.4 6.7 6.9 1.3
5b
6
6
6
9.4
8.6
8.4
1.2 78.0 8.3 7.6 3.2
1.5 78.7 9.2 7.9 2.2
0.6 82.0 9.8 5.3 0.8
6 b
7 b
40
8
80
8
(a) Mixed overnight at RT. (b) Reactions conducted using at least 1 g of 4. (c) Reaction conducted using 95 mg of 4.
The temperature dependence on formation of the cis impurity 36 was explored through a brief
computational study. The ground state energies of the cis and trans macrocycle products were
determined. The structures for 35 and 36 were optimized using DFT (B3 LYP-D3/6-31G*), and
the solvated ground state energies with thermal corrections at 25/40/80 °C were calculated at the
M06-2X/6-31G**/SM8(THF) level of theory (see the Supporting Information for details).30 It
was determined that at each temperature, the cis isomer had lower ground state energy by
approximately 4.1 kcal/mol. This result, taken in conjunction with the observed trans/cis ratio
dependence on temperature suggests the RCM reaction is operating under kinetic control;
increasing the reaction temperature provides access to the pathway leading to the cis impurity. In
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the optimized structures shown in Figure 2, the trans macrocycle has a “kink” in the backbone
when compared to the cis isomer, especially noticeable in the positioning of the quinoxaline ring
with respect to the backbone of the macrocycle. It is likely this feature that results in the
increased ground-state energy of the trans macrocycle compared to the more open cis isomer.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Left: DFT optimized structure of trans 35. Right: DFT optimized structure of cis 36.
A number of learnings from the RCM reaction in the ABT-450 process were implemented for
this substrate, especially with regard to methods for minimizing formation of dimer impurities
3
1
and driving the reaction to completion. The reaction was conducted at high dilution (80 mL
toluene per gram of 4, 0.020 M) to minimize formation of the dimer impurity 37. At higher
concentrations, the reaction initially produced more dimer 37 which slowly converted to product
3
5, but the final impurity profile was worse due to higher levels of the dimer impurity 37 and the
ligand derived impurity 38 (Table 4, entries 5 and 6). With the exception of entry 4, the reactions
in Table 4 were conducted with the simultaneous addition of separate solutions of catalyst and 4
over the same time period. The addition time of the solutions was approximately 6 hours but
ACS Paragon Plus Environment

Page 25 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
ranged from 4 to 8 hours (Table 4). The slow addition of substrate to the large volume of toluene
was intended to reduce the concentration of 4, and therefore minimize dimer impurity 37
formation in favor of the desired product 35. The slow addition of catalyst was intended to
ensure active catalyst was present throughout the addition. The RCM reaction produced ethylene
gas as a byproduct. Nitrogen sparging of the reaction mixture was used to remove the ethylene
and drive the reaction to completion. A nitrogen sparge rate of 10-12 standard cubic feet per
minute was implemented for scale up batches (Table 5).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
With the preferred reaction conditions shown in entry 7 of Table 4, the product 35 also
contained 10% of the cis isomer impurity 36, 5% of the dimer impurity 37, and 1-2% of both the
starting material 4 and the ligand impurity 38, along with 8 mol % of the catalyst. All of these
contaminants had to be removed by the workup and purification or removed in the downstream
processing. Crystallization of the macrocycle ester 35 was not effective for removal of the cis
isomer impurity 36, and instead crystallization of the next downstream intermediate, macrocyclic
acid 3, was determined to be the best point to reject the cis isomer impurity and other minor
impurities. The dimer impurity 37 and its daughter impurities formed downstream in the
synthesis of 3 and 1 were not effectively removed by crystallization. The method that completely
removed the dimer impurity 37 with acceptable product loss was silica gel chromatography of
intermediate 35. The silica gel purification also removed the RCM catalyst and the ligand
impurity 38. After the reaction was complete, imidazole was added to quench the reaction. Filtrol
was also added to the reaction mixture to adsorb the catalyst and improve the subsequent
filtration rate. The first silica gel plug (2 g silica gel/g 4) removed most of the catalyst, and the
second silica gel plug (6-7 g silica gel/g 4) removed the dimer impurity 37 and reduced the
ruthenium level down to ~10 ppm. The amount of silica gel and the solvent gradient for the
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
second plug filtration were selected to minimize the amount required while ensuring complete
removal of the dimer impurity 37.
The first large scale RCM reactions (Table 5, runs 1 and 2) were conducted using 80 volumes
of toluene at 40 °C, 8 mol % of catalyst, and a 6 hour addition of substrate and catalyst which
consistently gave the best results in terms of yield, reaction time, and impurity profile. Further
development for subsequent large scale batches showed that extending the addition time to 18 h
allowed the reaction to be conducted at 40 volumes of toluene and produced a modest increase in
yield while achieving a similar purity level post chromatography (Table 5, runs 3 and 4).
Table 5. Summary of Large-Scale RCM Reaction Results
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
F
F
F
F
N
N
N
N
O
O
O
O
Zhan 1B
toluene
O
O
N
OMe
N
OMe
O
O
40 °C
O
O
O
O
N
H
N
H
4
35
Parameters
Amount of 4 (kg)
Run 1 Run 2 Run 3 Run 4
19.7
80
19.7
80
20.0
40
20.0
40
Toluene (L/kg of 4)
Addition Time (h)
6
6
18
18
Product 35 (area %)
Cis Isomer 36 (area %)
Starting Material 4 (area %)
Assay yield %
87.2
10.5
2.3
86.0
10.4
3.6
88.7
10.6
0.5
88.1
10.8
1.0
70.7
64.8
71.2
75.9
ACS Paragon Plus Environment

Page 27 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The conversion of the macrocycle ester 35 to the macrocyclic acid 3, shown in Table 6, was
straightforward and proved to be the best point in the process for rejection of the cis isomer
impurity 39. The reaction produced the expected 10% of the cis isomer impurity 39, and after 4
crystallizations from 2-MeTHF and heptanes, the level of 39 was reduced to 1%. Although the
solvent usage for four crystallizations was higher than desired, the process was simple, and the
losses were modest as shown in Table 6. The overall yield for the conversion of the assayed
amount the macrocycle ester 35 to the macrocyclic acid 3 was 84% which was remarkable given
the degree of purification achieved.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 6. Saponification with Rejection of cis Isomer 39
F
F
F
F
N
N
N
N
O
O
O
O
aq. NaOH
O
O
N
OMe
N
OH
O
O
2-MeTHF
O
O
MeOH
O
O
N
H
N
H
84%
3
5 = trans
3 = trans
39 = cis
36 = cis
3
39
(area %)
Loss of 3
Sample
(
area %)
(%)
Post Workup
85.4
93.4
96.6
97.9
99.0
10.5
6.1
3.4
2.1
1.0
0.8
2.5
1.3
1.3
1.5
1
st Crystallization
2
nd Crystallization
rd Crystallization
th Crystallization
3
4
The synthesis of macrocyclic acid 3 required three chromatographic purification steps and four
crystallizations for a three-step yield of 57%. A more than 50% increase in yield for the enabling
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
process was achieved compared to the 35% yield reported for the same sequence in the initial
1
0
process. While the yield for the three step sequence to the macrocyclic acid 3 was acceptable
and the process could be used to generate multi-kilogram batches, the process would be
challenging for large scale production. The high RCM catalyst load and the reliance on silica gel
chromatography would be prohibitive from a cost and throughput perspective. Additionally, the
numerous crystallizations required to reduce the level of the cis isomer impurity 39 were also
prohibitive. It was clear that although the RCM route was capable of producing macrocyclic acid
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
, a more efficient route to the macrocycle would be required for large scale production.9
Final Assembly of Glecaprevir
The final coupling of amino sulfonamide 2 with macrocyclic acid 3 was accomplished using
EDAC and HOPO as the coupling reagents, as shown in Scheme 13. The coupling proceeded
cleanly without detectable epimerization of the proline carboxylate. The reaction was initially
run in dichloromethane with an aqueous workup to remove the reagent byproducts. The workup
was laborious and offered negligible impurity removal. The reaction was found to proceed
equally well in acetonitrile. The reaction was conducted in two reactors wherein the HOPO
active ester of macrocyclic acid 3 was prepared in one reactor and then slowly added to the other
reactor containing a mixture of amino sulfonamide 2 and triethylamine. Nearly complete
conversion of macrocyclic acid 3 was achieved using this order of addition with only 1.1
equivalents of amino sulfonamide 2.
ACS Paragon Plus Environment

Page 29 of 58
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Scheme 13. Final Coupling to Form Glecaprevir
F
F
F
F
N
N
N
N
O
O
O
O
1
) EDAC, HOPO,
O
O
MeCN, TEA
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
O
H
O
O
O
O
H2N
HCl
S
N
N
OH
N
S
+
N
H
O
2) MeOH/water
recrystallization
O
N
H
O
O
O
O
N
H
F
N
H
F
89%
F
F
3
2
1
The switch to acetonitrile as the reaction solvent also offered the opportunity for direct
crystallization after reaction completion. Upon quenching with aqueous acetic acid, the crude
glecaprevir (1) crystallized from the reaction mixture. The addition of acetic acid was necessary
to ensure protonation of the sulfonamide nitrogen and facilitate complete desupersaturation. The
crystallization of the crude API was accomplished in 93% yield with rejection of the cis isomer
daughter impurity to 0.37% from a starting level of 1.0% in macrocyclic acid 3. The crude API
was then recrystallized from methanol and water to obtain the mixed methanol-water solvate,
which then was dried under vacuum to remove methanol followed by humidification to yield the
3
2
desired hydrate crystal form. The overall yield through the two crystallizations was 89%, with
further rejection of the cis isomer daughter impurity to 0.15%. The process was used to prepare
nine batches of glecaprevir (1) totaling 41 kg.
Conclusion
An enabling synthesis of glecaprevir (1) was developed to support the pre-clinical evaluation
and subsequent Phase I clinical trials. The enabling route to glecaprevir (1) was established
through further development of the medicinal chemistry route. The development of the
fluorination of aldehyde 9 resulted in a reliable process suitable to meet the short term material
requirements of the program, along with a better understanding of impurities generated in the
ACS Paragon Plus Environment