Toward a Scalable Synthesis and Process for EMA401, Part I: Late Stage Process Development, Route Scouting, and ICH M7 Assessment

Article abstract of DOI:10.1021/acs.oprd.0c00215

We present the enantioselective synthesis of sodium (3S)-5-(benzyloxy)-2-(diphenylacetyl)-6-methoxy-1,2,3,4-tetrahydroisoquinoline-3-carboxylate (EMA401, olodanrigan), an angiotensin II type 2 antagonist. The manuscript features the process optimizations of the end game used for late phase clinical supplies, an overview of synthetic strategies identified in a route scouting exercise to a key intermediate phenylalanine derivative, and the analytical control strategy of the potentially formed highly toxic impurity bis(chloromethyl) ether (BCME). Starting from the phenylalanine derivative, we describe the optimizations of the end game from early phase to late phase processes with consequent improvements in the PMI factor. This sequence includes a Pictet-Spengler cyclization and an amide coupling as the last bond-forming steps, and the manufacturing process was successfully implemented on a 175 kg scale in a pilot plant setup. The modified process conditions eliminated one step by in situ activation of the carboxylic acid, avoided the REACH listed solvent DMF, and resulted in a PMI improvement by a factor of 3. In the final crystallization, a new, thermodynamically more stable modification of the drug substance was found in the complex solid-state landscape of EMA401 during an extensive polymorph screening. A process suitable for large-scale production was developed to prepare the new polymorph, avoiding the need of any special equipment such as fluidized bed drying required in the early phase process. In the second section, some of the synthetic approaches investigated for the route scouting of the phenylalanine derivative key intermediate are presented. To conclude, we discuss the analytical control strategy for BCME, the formation of which, due to the simultaneous presence of HCl and CH2O in the Pictet-Spengler cyclization, could not be ruled out. The BCME purge factor calculations using the tools of ICH M7 control option 4 are compared to actual results from spiking experiments.

Full text of DOI:10.1021/acs.oprd.0c00215

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Full Paper
Towards a Scalable Synthesis and Process for EMA401. Part I: Late
Stage Process Development, Route Scouting and ICH M7 Assessment
Leo Albert Hardegger, Franck Mallet, Barbara Bianchi, Chunlong Cai, Alexandre Grand-Guillaume
Perrenoud, Roger Humair, Richard Kaehny, Stephan Lanz, Cheng Li, Jialiang Li, Florian A.
Rampf, Lei Shi, Christoph Spoendlin, Jeannine Staeuble, Shangjun Teng, Xiangguang Tian,
Bernhard Wietfeld, Yao Yang, Bo Yu, Christine Zepperitz, Xuesong Zhang, and Yong Zhang
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.0c00215 • Publication Date (Web): 20 Aug 2020
Downloaded from pubs.acs.org on August 20, 2020
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 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Towards a Scalable Synthesis and Process for
EMA401. Part I: Late Stage Process Development,
Route Scouting and ICH M7 Assessment
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
†
¥
†1
‡2
Leo A. Hardegger *, Franck Mallet , Barbara Bianchi , Chunlong Cai , Alexandre
†
†
†
£
Grand-Guillaume Perrenoud , Roger Humair , Richard Kaehny , Stephan Lanz , Cheng
‡3
‡
†
‡
†
¥
Li , Jialiang Li , Florian Rampf , Lei Shi , Christoph Spoendlin , Jeannine Stäuble ,
‡4
‡
†
‡
‡
Shangjun Teng , Xiangguang Tian , Bernhard Wietfeld , Yao Yang , Bo Yu , Christine
†5
‡6
‡7
Zepperitz , Xuesong Zhang , Yong Zhang
†Chemical and Analytical Development, Novartis Pharma AG, 4002 Basel, Switzerland
¥Pharmaceutical and Analytical Development, Novartis Pharma AG, 4002 Basel,
Switzerland
‡Suzhou Novartis Technical Development Co., Ltd., Changshu City, 215537, China
£Interlabor Belp AG, 3123 Belp, Switzerland
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 62
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
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
TOC
Cl
O
Cl
ICH M7 assessment
BCME
?
?
HO
O
HCl
OBn
O
OBn
O
O
O
?
CH2O
O
O
O
(S)
OH
(S)
OH
(S)
O Na
O
Route scouting
NH2
NH
N
?
?
Late-stage process development
EMA401 / olodanrigan
ACS Paragon Plus Environment
2

Page 3 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
ABSTRACT We present the enantioselective synthesis of sodium (3S)-5-(benzyloxy)-
2
-(diphenylacetyl)-6-methoxy-1,2,3,4-tetrahydroisoquinoline-3-carboxylate
(EMA401,
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
olodanrigan), an angiotensin II type 2 antagonist. The manuscript features the process
optimizations of the end game used for late phase clinical supplies; an overview of
synthetic strategies identified in a route scouting exercise to a key intermediate
phenylalanine derivative; and the analytical control strategy of the potentially formed
highly toxic impurity bis(chloromethyl)ether (BCME). Starting from the phenylalanine
derivative, we describe the optimizations of the end game from early phase to late
phase processes, with consequent improvements in the PMI factor. This sequence
includes a Pictet-Spengler cyclization and an amide coupling as the last bond-forming
steps, and the manufacturing process was successfully implemented on a 175 kg scale
in a pilot plant setup. The modified process conditions eliminated one step by in situ
activation of the carboxylic acid, avoided the REACH listed solvent DMF, and resulted in
a PMI improvement by a factor of 3. In the final crystallization, a new,
thermodynamically more stable, modification of the drug substance was found in the
complex solid-state landscape of EMA401 during an extensive polymorph screening. A
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 62
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 suitable for large-scale production was developed to prepare the new
polymorph, avoiding the need of any special equipment such as fluidized-bed drying
required in the early phase process. In the second section, some of the synthetic
approaches investigated for the route scouting of the phenylalanine derivative key
intermediate are presented. To conclude, we discuss the analytical control strategy for
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
BCME, the formation of which, due to the simultaneous presence of HCl and CH O in
2
the Pictet-Spengler cyclization, could not be ruled out. The BCME purge factor
calculations using the tools of ICH M7 control option 4 are compared to actual results
from spiking experiments.
KEYWORDS EMA401, olodanrigan, phenylalanine derivatives, scale-up, route
scouting, ICH M7 control option 4, BCME, bis(chloromethyl)ether
Introduction
The first report of EMA401 (1; olodanrigan) dates back to the early 1990s,1,2 when the
structure of EMA401 (1) was identified in a program aiming at the development of
ACS Paragon Plus Environment
4

Page 5 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
angiotensin II type 2 (AT R) antagonists. Twenty years after its first disclosure, the
2
potential of EMA401 as a AT R antagonist in the treatment for neuropathic pain was
2
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
recognized and the results of clinical studies were published,3–6 and in 2015 the
compound was in-licensed to Novartis. When we started process development work on
EMA401 (1) as its sodium salt, we performed both i) an extensive route scouting
exercise to find an enantioselective, environmentally-friendly and cost-effective
synthesis of EMA401, and ii) ran a complete program to find a new, more stable
polymorph of the drug substance. Herein, we report our results from the process
optimizations for the synthetic end game and the final crystallization, evaluate different
synthetic strategies to the phenylalanine key derivative, and discuss the comparison
between ICH M7 control option 4 with spiking experiments for bis(chloromethyl)ether
(BCME), a highly toxic, potential impurity of the penultimate bond-forming step of the
EMA401 synthesis.
Results and Discussion
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 62
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
Late stage synthesis of EMA401 (1)
The retrosynthetic analysis of EMA401 (1) provides a number of potential
disconnections as late stage conversions. Based on chemical and technical feasibility,
strategic considerations, health, safety and environment (HSE) assessments, and cost
calculations, the synthetic end game depicted in Scheme 1 for EMA401 (1) was
selected. The same synthetic route had already been used for early phase development
and in all syntheses published on EMA401 (1),1–3,6,7 which allowed us to freeze the
impurity profile on the drug substance. Starting from phenylalanine derivative 2,
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
tetrahydroisoquinoline
3
was synthesized in
a
Pictet-Spengler cyclization.
Tetrahydroisoquinoline 3 was coupled in the last bond-forming step to diphenylacetic
acid (4) to give carboxylic acid [5].
HO
O
OH
4
OBn
O
OBn
O
O
O
O
O
O
O
O
O
CH2O
NaOMe
O
O
O
O
O
O
(S)
NH2
OH
(S)
OH
(S)
OH
O
(S)
O
O
Na
(S)
O
O
Na
(S)
O
O
Na
iPrOH
EtOAc
NH
N
N
N
N
2
3
5
6
7
1
Scheme 1. Manufacturing route of the synthetic end game of EMA401 (1). Starting from
phenylalanine derivative 2, EMA401 (1) was synthesized in a Pictet-Spengler cyclization
ACS Paragon Plus Environment
6

Page 7 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
to afford tetrahydroisoquinoline 3, and subsequently coupled with diphenylacetic acid
(4).
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
Addition of NaOMe to carboxylic acid [5] gave carboxylate [6], which was isolated as
its iPrOH solvate 7. The final crystallization afforded EMA401 drug substance (1) as a
sodium salt. An in-depth route scouting exercise conducted for the synthetic approach
to the key intermediate phenylalanine derivative 2 is presented in a separate Section in
this manuscript (vide infra).
Process Development of the Pictet-Spengler Cyclization to Tetrahydroisoquinoline 3
The early phase process for the Pictet-Spengler cyclization from phenylalanine
derivative 2, used as its hydrochloride salt and acetic acid solvate, to
6
tetrahydroisoquinoline 3 is given in Scheme 2. The reaction was run in an aqueous
phosphate buffer system at pH 2. An initial assessment of the process identified the
following challenges: i) long reaction times of 18–32 h, ii) a low concentration (0.1 M), iii)
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 62
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
a slow filtration during the isolation of tetrahydroisoquinoline 3, iv) the product had a
tendency to float on the aqueous layer on scale, causing IPCs to fail due to non-
representative sampling, and v) an assessment of the commercial production
capabilities identified this step as the bottle-neck defining the cycle time of the entire
drug substance production.
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
HCl
AcOH
30% Formaline (1.7 eq)
O
O
O
O
aq. H PO /NaH PO buffer
pH = 2
3
4
2
4
O
O
OH
OH
6
0 °C, 18-32 h,
then
NaOAc to pH = 3.3
filter, wash with H O & acetone
NH2
NH
2
2
3
93% isolated yield
Scheme 2. Early phase process for the Pictet-Spengler reaction to
tetrahydroisoquinoline 3 in an aqueous buffer.
Attempts to decrease the reaction time by increasing the temperature or by
2
decreasing the pH led to the partial cleavage of the benzyl group. While the thus
formed phenol 8 (Scheme 3) could be purged in the isolation of 3 as its zwitter ion, its
formation led to a significant reduction of the yield. We therefore turned our attention to
the addition of co-solvents aiming to increase the concentration and to decrease the
ACS Paragon Plus Environment
8

Page 9 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
reaction time. However, none of the co-solvents screened showed an improvement in
terms of space-time-yield. Gratifyingly when we ran the cyclization entirely in acetic acid
as solvent, full conversion was observed at 50 °C within 1 h in a slurry-to-slurry process
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). In addition, the concentration could be increased to 0.5 M (5 v/w with
respect to phenylalanine 2) and the equivalents of para-formaldehyde could be reduced
to 1.1 eq. As phenylalanine 2 was isolated as its hydrochloride salt,
tetrahydroisoquinoline 3 was also isolated as its hydrochloride salt after cooling to 20
°C, with much improved filtration properties as compared to the zwitter ion of 3. The
filter cake was then washed with acetone to purge acetic acid, residual phenylalanine 2,
and the three main impurities (Scheme 3). The de-benzylated phenol 8 originated
mostly from the synthesis of phenylalanine derivative 2, and had to be analytically
controlled by specification on phenylalanine 2. While phenol 8 was purged in the
isolation of tetrahydroisoquinoline 3 as the zwitter ion, purging of de-benzylated impurity
8 in the isolation of 3 as the hydrochloride salt and in the subsequent steps to the drug
substance was low. The N-acetyl byproduct 9 (Scheme 3) was observed at levels up to
0.38a% in stress tests and after spiking with acetic anhydride, and was purged in the
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 62
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
isolation of 3·HCl. Lactone byproduct 10 was observed at less than 1a% in IPCs and
purged to below limit of detection (<LOD) in the isolation of 3·HCl. Finally, the water
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
content in the reaction was found to be critical and the addition of 2wt% H O with
2
respect to phenylalanine 2 caused the loss of more than 10% tetrahydroisoquinoline 3
in the mother liquor, due to the increased solubility of 3 in aqueous acidic conditions.
The residual formaldehyde content in the isolated 3·HCl was determined to be 77
ppm, i.e. already by a factor of more than 200 below the permitted content in the drug
substance for oral administration (oral limit for formaldehyde : maximum daily dose of
8
EMA401; 10’000 g/d : 0.6 g/d = 16’666 ppm). None of the safety measurements
indicated a critical exotherm for the reaction, reagents or product. In summary, the
space-time-yield could be improved by a factor of 20, the number of unit operations as
well as the number of discrete raw materials required was significantly reduced as
compared to the early phase process, which led to a reduction of the PMI by a factor of
5 (Table 1). The modified process was successfully implemented on 1 kg of
phenylalanine derivative 2·HCl. It is noted that the process in aqueous medium as well
ACS Paragon Plus Environment
10

Page 11 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
as the process in acetic acid could potentially lead to the formation of the theoretical
impurity bis(chloromethyl)ether (BCME), a highly toxic by-product (vide infra).
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
HCl
O
AcOH
OBn
OH
(CH2O)n (1.1 eq)
O
O
O
O
OH
O
O
O
AcOH,
50 °C, 1 h,
then
O
O
O
O
O
OH
OH
NH HCl
OH
OH
O
O
NH2
NH
N
filter and wash with aceton
92% isolated yield
OBn
2
3
8
9
O
10
Scheme 3. Modified process of the Pictet-Spengler reaction to tetrahydroisoquinoline
3·HCl in acetic acid. By-products phenol 8, N-acetyl derivative 9 and lactone 10 were
observed in this step in IPCs and the isolated product.
Table 1. Comparison of the original Pictet-Spengler processes to tetrahydroisoquinoline
3
in aqueous buffer and optimized process to 3·HCl in acetic acid.
Early phase process in aqueous buffer
Long reaction times: 18–32 h
Low concentration: 0.1 M
Modified process in acetic acid
Short reaction times: 1 h
High concentration: 0.5 M
Fast filtration of 3·HCl
Slow filtration of 3
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 62
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
Floating of 3 on aqueous mixture on scale No floating observed, slurry to slurry
reaction
Increasing temperature or lowering pH led Extended reaction time led to by-products
to the formation of by-products
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
PMI = 55
PMI = 11
Amide coupling and salt formation to intermediate 7
The early phase process for the amide coupling to intermediate [5] from 3 and 4 is
3
depicted in Scheme 4. Diphenylacetic acid (4) was activated as its acyl pyrrazole
derivative 11, a stable compound isolated after several washing steps. Pyrrazole 11
was then added to a mixture of 3 and tetramethylguanidine in DMF. After complete
conversion to amide [5], toluene was added and the mixture was subjected to an
aqueous acidic quench and aqueous washings, followed by the addition of NaOMe to
obtain carboxylate [6], a solvent switch to EtOAc, and finally to the crystallization as the
iPrOH solvate 7. We aimed at streamlining this sequence of processes to i) reduce the
amount and number of chemicals used, and hence the amount of waste created, ii)
reduce the number of operations in each process, iii) avoid the isolation of pyrrazole 11,
ACS Paragon Plus Environment
12

Page 13 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
and iv) replace the REACH listed solvent DMF to isolate the isopropanol solvate 7 in
high yield with a comparable impurity profile.
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
H
N
N
N
O
OH
N
O
CDI
EtOAc,
several aq. washes
4
11
OH
NH
N
N
N
N
O
O
O
DMF
then
O
O
O
ONa
O
1) toluene, aq. washes
) NaOMe in MeOH
3) Switch to EtOAc
) iPrOH
N
O
OH
2
NH
4
3
11
7
Scheme 4. Early phase synthesis of iPrOH solvate 7 from tetrahydroisoquinoline 3 and
pyrrazole derivative 11.3
We aimed at the in-situ activation of diphenylacetic acid (4) to avoid the isolation of
pyrrazole derivative 11. Since both, 3 and 4 bear a carboxylic acid functionality, it was of
obvious importance to activate only the carboxylic acid on diphenylacetic acid (4).
Activation of the carboxylic acid on tetrahydroisoquinoline 3 led to the erosion of the
chiral purity and to the formation of coupling products 12 and 13 (Figure 1). These
ACS Paragon Plus Environment
13

Organic Process Research & Development
Page 14 of 62
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
impurities could be purged in the crystallization of 7, albeit at the cost of a significantly
reduced yield.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
O
O
O
O
HO
O
N
O
N
N
O
O
O
O
O
N
12
13
Figure 1. Impurities 12 and 13 were observed when the carboxylic acid on
tetrahydroisoquinoline 3 was inadvertently activated by an excess of activating reagent.
Since EMA401 was predicted to be a multi ton project, we initiated a screening for
amide coupling conditions with a special focus on the large scale availability and price
of the amide coupling reagent as recently discussed in this journal.10 The obvious
choice would have been Schotten-Baumann conditions employing, for example, the
acid chloride of diphenylacetic acid. In initial screenings, those conditions led to an
increased number of impurities, and to an erosion of the chiral purity as noted in earlier
2
reports. We therefore decided to look for alternative conditions to deliver material for
ACS Paragon Plus Environment
14

Page 15 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
clinical studies and focused our screening on surfactants, which had shown to be an
11
excellent alternative to dipolar aprotic solvents for amide coupling reactions. To our
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
dismay, none of the conditions screened gave full and clean conversion, most likely due
to solubility issues observed with tetrahydroisoquinoline 3. In a next phase, we
screened amide coupling reagents in organic solvents. Of all reagents tested, 1,2-
carbonyldiimidazole (CDI) showed the most promising results. Most other reagents
tested (PivCl, isobutyl chloroformate, T3P, EDC, DCC and triazine-based reagents)
suffered from poor conversion, formation of impurities, or required the use of HSE
critical additives such as HOBt for EDC and DCC.
The low solubility of tetrahydroisoquinoline 3 in most organic solvents mandated the
use of a base to increase its solubility. Various bases were screened (DBU, Na CO , Na
2
3
imidazolate, NMM, TMG, DIPEA) and out of those only DBU was capable of forming a
stirrable suspension of tetrahydroisoquinoline 3 in organic solvents, resulting in full
conversion to [5] with a low by-product profile. Since DBU is a strong base, which
potentially could also lead to E cb elimination of the activated diphenylacetic acid (4),
1
the equivalents were kept at below 1.1 with respect to tetrahydroisoquinoline 3.
ACS Paragon Plus Environment
15

Organic Process Research & Development
Page 16 of 62
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
Less polar solvents such as THF, MeTHF, iPrOAc, acetone, or toluene led to the
formation of gels or unstirrable suspensions of 3, even when DBU was used as a base.
From the existing process, we knew that DMF afforded a well-stirrable suspension of 3.
This behavior was also found with 1-methylpyrrolidin-2-one (NMP, CAS# 872-50-4), 1-
butylpyrrolidin-2-one (NBP, CAS# 3470-98-2), 1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-
pyrimidinone (DMPU, CAS# 7226-23-5) and 4-formylmorpholine (NFM, CAS# 4394-85-
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
8). Combined with the fact that the sodium salt [6] was soluble in toluene, we focused
our attention to solvent mixtures of dipolar aprotic solvents with toluene. From the four
dipolar aprotic solvents above, NFM was found to be the most benign option in terms of
HSE and compatibility with downstream processing, and further optimization was
continued with 4:1 (v/v) mixtures of toluene/NFM as the solvent. The ratio was later
optimized to 3.25:1 (w/w). Using less NFM led to the formation of biphasic mixtures
resulting in slower conversion and an increased impurity profile.
As discussed above, the stoichiometry of CDI was critical to avoid the formation of
impurities. Additional CDI could not be dosed once the solution of imidazole derivative
[14] had been added to tetrahydroisoquinoline 3 (Scheme 5), therefore, specifications
ACS Paragon Plus Environment
16

Page 17 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
for CDI and the implementation of an IPC for the activation of diphenylacetic acid (4)
were crucial for the entire process. Hence, a solution of 4 (1.20 eq, i.e. an excess as
compared to CDI) in toluene/NFM 4:1 was added to a suspension of CDI (1.15 eq) in
toluene/NFM 4:1. At the end of the addition, a slightly yellow, clear solution of imidazole
derivative [14] was obtained, which was added to a suspension of 3 in toluene/NFM 4:1
with 1.06 eq (2.06 eq) of DBU when starting from the zwitter ion of 3 (the hydrochloride
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·HCl). Complete conversion of tetrahydroisoquinoline 3 was achieved within 3 h at 30
°C, indicated by the change in aspect from a suspension to a solution.
In the early phase process, the post-reaction quench of an excess of pyrrazole 11 with
dimethylethylenediamine (CAS# 108-00-9) was routinely done in all GMP campaigns.
Lab results indicated that with the change from pyrrazole 11 to the less stable imidazole
[14] in the late phase process, the addition of dimethylethylenediamine might not be
required for the removal of by-products originating from diphenylacetic acid, but that it
may act as a buffer in the subsequent quench with H SO to maintain the chiral purity.
2
4
Indeed this first H SO quench proved to be pivotal for the chiral purity, with stirring
2
4
speed, addition time and temperature all having an influence on the chiral purity of [5].12
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 62
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
To extract all amines present in the mixture, a total of two acidic and two neutral washes
were implemented.
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
Lab experiments and IPCs on scale indicated that NFM mostly hydrolyzed under the
aqueous acidic conditions with 1/2 = 25 min at 25 °C (see Supporting Information (SI))
and only small amounts of NFM remained in the toluene layer after four aqueous
washes. The toluene solution of carboxylic acid [5] was treated with a solution of
NaOMe (1.0 eq) in MeOH to obtain sodium salt [6]. When more than 1.2 eq. of NaOMe
were used, the excess of NaOMe could react with traces of NFM and water in the
toluene phase, leading to the formation of sodium formate, which was not purged and
carried through as an impurity to the drug substance.
Azeotropic removal of H O and MeOH from the solution of [6] in toluene, and
2
subsequent dilution with toluene afforded a solution of [6], which was then added to
iPrOH at 50 °C, resulting in the formation of the desired iPrOH solvate 7. Addition at
lower temperatures did not afford the iPrOH solvate, but upon heating the mixtures to
above 45 °C, the formation of 7 as a white precipitate was observed.
ACS Paragon Plus Environment
18

Page 19 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Both addition orders, addition of the toluene solution of [6] to iPrOH and iPrOH to the
toluene solution of [6], were suitable. When the solution of [6] in toluene was added to
iPrOH, a suspension with superior stirring properties and a more controllable
crystallization was achieved. Finally, the suspension of 7 was cooled, filtered and
washed with iPrOH to afford intermediate 7 in around 90% yield, and a purity greater
than 99.6a%, with diphenylacetic acid (4) being the only impurity detected. Experiments
indicated that in the subsequent step to the drug substance 1, diphenylacetic acid (4)
could be depleted from 1.2wt% in intermediate 7 to <LOD in the drug substance. The
modified process was successfully run at a 175 kg scale of tetrahydroisoquinoline 3.
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
1.15 eq
N
O
O
Na
N
O
OH
N
N
O
N
NFM (CAS# 4394-85-8)
H
O
N
O
O
OH
O
N
toluene/NMF 4:1
N
O
4
[14]
7
1.20 eq
iPrOH
Cl
O
Cl
N
NH2
0% H2SO4
0% H2SO4
H2O
H2O
BCME
1
.06 eq
1
1
N
Na
O
O
O
O
N
O
O
O
O
OH
O
O
toluene/NMF 4:1
NaOMe (1.0 eq)
O
N
N
O
OH
NH
3
[5]
[6]
1.0 eq
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 62
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 5. Process to iPrOH solvate 7 with activation of diphenylacetic acid (4) with
CDI and telescoping of the corresponding imidazole derivative [14] into the amide
coupling step to carboxylic acid [5]. For BCME spiking, see the last Section of this
manuscript.
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 2. Comparison of the early and late phase processes to iPrOH solvate 7.
Early phase process to 7
Two isolated steps
Late phase process to 7
One isolated step
Combined reaction time: 21 h
Combined reaction time: 4 h
Many aqueous washes in the isolation of Telescoping of imidazole derivative [14]
pyrrazole derivative 11
DMF (ICH class 2 solvent) used in DMF replaced by 4-formylmorpholine
penultimate step
(NFM)
PMI = 89
PMI = 32
Crystallization of EMA401 (1) drug substance
EMA401 as its sodium salt has a highly complex solid-state landscape (Figure SI2).
The sodium salt of EMA401 (1) can crystallize in different solid states as an isopropanol
solvate, a mixed solvate MeOH-EtOAc, a mesomorph produced via desolvation of the
ACS Paragon Plus Environment
20

Page 21 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
isopropanol solvate, several hydrates, amorphous and as an anhydrous phase. The
latter was our target phase, which was stable up to 65% relative humidity at 25 °C, a
limit applied to all the steps from production to downstream processing.
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
A targeted solvent screen was carried out for EMA401 (1) in order to obtain the
anhydrous phase (modification B, Figure SI2). The solvent system selected was
water/EtOAc, even though several hydrates were found. The rationale behind this
decision was that due to the existence of several hydrate phases of EMA401 (1), the
addition of water to EtOAc would lead to an increase in solubility. Figure 2 shows the
temperature-dependent solubility of EMA401 in mixtures of water and EtOAc.
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 62
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
Figure 2. Solubility of EMA401 (1) sodium salt as a function of temperature and water
content in EtOAc. Solubility values were determined gravimetrically and are given in
w/w of solution.
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
EMA401 exhibits a retrograde solubility, i.e. the solubility decreases at higher
temperature. The temperature-dependent solubility was found to be less pronounced
than the change with the content of water added to the solvent system. Hence, the
crystallization procedure was developed as a seeded crystallization by antisolvent
addition, i.e. EtOAc was added to reduce the water concentration to less than 1wt% as
the final concentration. Following this crystallization procedure, EMA401 (1) was
obtained as crystalline needles (see Experimental Details and Figure SI3).
The crystallization procedure was successfully scaled up to 90 kg scale. As the
procedure was an antisolvent crystallization, the final volume was high and an
alternative procedure including a constant volume distillation was developed in order to
increase the productivity and successfully scaled up to 90 kg. For both procedures, a
final wash with heptane was implemented, as the filter cake was hygroscopic when
ACS Paragon Plus Environment
22

Page 23 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
wetted with EtOAc, and therefore would have induced issues when unloading the wet
cake from the filter to the tray dryer, a common equipment in pilot plants. This additional
wash was not required when working with a filter dryer, as the product would remain
under dry nitrogen. The XRPD (Figure SI4) showed the evolution of a wet cake of
EMA401 (1) under ambient conditions. One of the hydrates was detected after a few
hours and led to the formation of a sticky material (hygroscopicity of the solid) and
concomitant loss of crystallinity, if water-content was not completely controlled.
Importantly, the modified process to the new polymorph no longer required special
equipment such as fluidized-bed drying required for the early phase process.
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
Route scouting for the key intermediate phenylalanine derivative 2
The synthesis of optically pure non-natural amino acid derivatives remains a synthetic
challenge, in particular for process chemistry. For the synthetic access to the key
intermediate phenylalanine derivative 2, we grouped the most promising approaches for
scale-up identified in an internal route scouting exercise as shown in Scheme 6–8. The
synthetic route used for the delivery of drug substance for early phase clinical trials
ACS Paragon Plus Environment
23

Organic Process Research & Development
Page 24 of 62
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
relied on the asymmetric hydrogenation approach developed by Burk et al. for N-acetyl
protected derivatives (Scheme 6).6,13 ortho-Vanillin (15), which is available on bulk scale
at high purity and low price, was treated with BnBr to afford benzylated ortho-vanillin 16,
which was subjected to a Horner-Wadsworth-Emmons reaction with phosphonate 17a,
to deliver enamide 18a in good yield and excellent selectivity. The key step of this route
was the asymmetric hydrogenation of enamide 18a to establish the chiral center in
amide 19a, and a screening of ligands, precatalysts and solvents indicated
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
[Rh(COD) ]OTf and the Josiphos ligand SL-J006-2 in THF or MeOH to deliver 19a with
2
full conversion, high selectivity and excellent chiral purity (Scheme 6). In our hands,
MeOH as solvent delivered a more robust reaction than in THF. The catalyst loading
could be reduced to 1:2’000 (w/w, C/S) and the asymmetric hydrogenation step was
scaled up to a maximum batch size of 110 kg of enamide 18a.
ACS Paragon Plus Environment
24

Page 25 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
O
P
EtO
OEt
OMe/Et
R
N
O
O
R
O
O
R
19a + c) i) Boc2O, DMAP
H
O
O
Me/Et
H2,
Rh cat
SL-J006-2
O
Me/Et
ii) LiOH
iii) HCl
O
O
1
1
1
7a: R = Ac
7b: R = Boc
7c: R = formyl
(S)
HN
HN
19b) HCl
19d) i) LiOH
18a: Ac, 93% with 17a
97% from 22a
18b: Boc, 94%
18c: Formyl, 55% with 20
18d: Benzoyl, 92%
19a: Ac
19b: Boc
19c
19d: Benzoyl
ii) HCl
O
C
:
Formyl
N
Et/MeO
20
OR
O
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
O
O
(
S)
OH
R = H
R = Bn, 97% 16
15
BnBr
R
N
H
OMe/Et
NH2
O
2, 81% (3 steps) from 18a
2
2
1a R = Me
1d R = Bz
lipase
reduction
O
O
O
O
P(C6H11)2
O
O
(S)
P(3,5-diCF3C6H3)2
O
O
N
N
Fe
22a: R = Me, ~60%
2d: R = Ph, ~80%
R
23
R
2
SL-J006-2
Scheme 6. Routes to phenylalanine derivative 2 via asymmetric hydrogenation as the
key step.
In order to cleave the N-acetyl group under mild conditions, a Boc-group had to be
introduced (Boc O, DMAP) on the nitrogen atom, followed by the cleavage of the acetyl
2
group and concomitant saponification of the ester under basic conditions (LiOH), and
subsequent carbamate deprotection under acidic conditions (HCl, AcOH) to obtain
6
phenylalanine derivative 2 as the HCl salt and AcOH solvate. N-acetyl deprotection
under acidic conditions at elevated temperature was not practical in our case due the
lability of the benzyl group under these conditions.2
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 62
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
Alternative approaches were investigated to avoid the functional group
interconversions, for example the use of Boc-protected didehydroamino acid derivative
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
18b (Scheme 6). However, the asymmetric hydrogenation with N-Boc derivative 18b
was not found to be robust, requiring higher catalyst loading, leading to stalled reactions
as well as incomplete conversions. The N-Boc derivative 18b was synthesized in a
Horner-Wadsworth-Emmons reaction from benzylated ortho-vanillin 16 and HWE
14
reagent 17b, which itself was prepared from the corresponding Cbz derivative. This
led to the conclusion that the overall gain in protecting group efficiency was negligible.
Another alternative target designed to overcome the functional group interconversion
15
was enamide 18c, the corresponding N-formyl protected derivative. The HWE reagent
17c required was prepared in analogy to similar reagents, but its isolation was
16
complicated due to its physical properties as a viscous oil. The alternative approach
via isocyanate 20 afforded mixtures of the E- and Z-isomer, with the desired E-isomer
isolated as a solid, while the Z-isomer was a viscous oil and purged to the mother liquor.
This led to low isolated yields of E-isomer 18c, while the mother liquor consistently
contained a 1:1 mixture of the E- and Z-isomers of 18c. In addition, isocyanates have
ACS Paragon Plus Environment
26

Page 27 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
unfavorable health, safety and environment (HSE) properties and in our case required
the removal of black impurities by distillation directly before their use in the subsequent
condensation reaction.
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 raw material costs of the HWE reagents 17a–c contributed significantly to the
overall drug substance costs, while only delivering five heavy atoms. In our attempts to
avoid the HWE reagent, we investigated the approach via the Erlenmeyer-Plöchl
17
azlactone condensation (Scheme 6). When starting from aceturic acid (21a), isolated
yields of azlactone 22a were usually moderate (around 60%). The subsequent
solvolysis to enamide 18a could then be achieved in good yields. The yield of the
Erlenmeyer-Plöchl condensation product azlactone 22d was improved to approx. 80%
when using hippuric acid (21d). However, the subsequent solvolysis to enamide 18d
was found to be capricious, and the removal of benzoic acid proved to be challenging.
An alternative approach to phenylalanine 2 from azlactones 22 would have been the
use of lipases after reduction to lactone 23. However, first screening hits advised
against this synthetic approach, also considering a maximum yield of 50% and the need
for re-cycling of the undesired enantiomer. Preparation of amino acid 2 via the formation
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 62
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
7
of coumarins or a nitro derivative, as described in prior literature, were not pursued, the
latter due to safety concerns. An alternative approach via a hydantoin intermediate had
also been disclosed previously.1,2,6 Access to amino acid via hydantoin lyase has not
been pursued in depth due to prior in-house knowledge advising against this enzyme
class.
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
We also aimed at developing a synthetic approach relying on the substitution of
bromide 24, obtained from benzylic alcohol 25, with protected glycine 26 (Scheme 7).1,18
This step was run under phase-transfer catalysis conditions, using asymmetric or
achiral phase-transfer catalysts. Various different cinchona-based and Maruoka-type
phase-transfer catalysts were screened. The best results were obtained with 0.5 mol%
of Maruoka-type phase-transfer catalyst 27 delivering amino acid 2 in 70% yield and
90% ee. Attempts to enrich the chiral purity by dynamic kinetic resolution of the racemic
amino acid 2, or ester and imine derivatives thereof, obtained by using an achiral
phase-transfer catalyst in the conversion of derivatives of 24 and 26, were not fruitful.
Equally the application of conglomerate crystallization failed to meet the requirements.
All these results let us to conclude that the synthetic approach depicted in Scheme 7
ACS Paragon Plus Environment
28

Page 29 of 62
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
was not the most suitable strategy to phenylalanine derivative 2 for scientific and
economic reasons.
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
F
F
F
O
N
O
OBn
OBn
OBn
1) Maruoka Catalyst
KOH,H2O/toluene
Br
N
O
NaBH4
O
PBr3
O
O
OH
Br
O
2) TBAB
O
CO2H
NH2
2, 70%, 90% ee
16
25, 97%
24, 90%
26
F
27
F
F
Scheme 7. Synthetic approach to phenylalanine derivative 2 using phase-transfer
catalysis
The aminotransferase-catalyzed amination of ketoacid 28 offered yet another
8).1,2,6
alternative approach towards phenylalanine derivative
2
(Scheme
Transaminases have become increasingly popular for large scale production as
19
highlighted by recent publications. We were pleased to find that in an initial screening
with ketoacid 28, we were able to detect conversion to phenylalanine 2. Ketoacid 28
was prepared in two synthetic steps from benzylated ortho-vanillin 16 via its hydantoin
1
derivative 29. The condensation with hydantoin could be achieved with 2-aminoethanol
ACS Paragon Plus Environment
29