Development of the Late-Phase Manufacturing Process of ZPL389: Control of Process Impurities by Enhanced Process Knowledge

Article abstract of DOI:10.1021/acs.oprd.1c00077

The development of the late-phase manufacturing process of the drug candidate ZPL389 and the strategies for the control of impurities are outlined in detail. Selective salt formation at several stages was pivotal to controlling the process impurities. The extensive optimization of the N-methylation of a Boc-protected amine with dimethyl sulfate and of a nucleophilic aromatic substitution without the use of metal catalysts led to a robust, scalable process. The process was demonstrated on a >100 kg scale. Overall, improved drug substance quality, higher yield, and reduction of the process mass intensity were achieved.

Full text of DOI:10.1021/acs.oprd.1c00077

pubs.acs.org/OPRD
Article
Development of the Late-Phase Manufacturing Process of ZPL389:
Control of Process Impurities by Enhanced Process Knowledge
Ernesto Santandrea,* Christine Waldraff, Gilles Gerber, Maël Moreau, and Pascal Beney
Cite This: https://doi.org/10.1021/acs.oprd.1c00077
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: The development of the late-phase manufacturing process of the drug candidate ZPL389 and the strategies for the
control of impurities are outlined in detail. Selective salt formation at several stages was pivotal to controlling the process impurities.
The extensive optimization of the N-methylation of a Boc-protected amine with dimethyl sulfate and of a nucleophilic aromatic
substitution without the use of metal catalysts led to a robust, scalable process. The process was demonstrated on a >100 kg scale.
Overall, improved drug substance quality, higher yield, and reduction of the process mass intensity were achieved.
KEYWORDS: ZPL389, 2-aminopyrimidines, salt formation, nucleophilic aromatic substitution, metal-free aromatic amination,
impurity control
characterization of process impurities and definition of
strategies to their control were paramount. Therefore, a
development program to address all these issues was initiated.
Herein, we describe the choice of regulatory starting materials,
the efforts at characterizing all relevant impurities, and the
development of all manufacturing steps.
INTRODUCTION
■
Atopic dermatitis is an inflammatory condition, whose
symptoms result in poor life quality for the patients affected
by it. ZPL389 (Figure 1), formerly also known as ZPL-
RESULTS AND DISCUSSION
■
Starting Material 1. 4,6-Dichloropyrimidine-2-amine (1)
was selected as the regulatory GMP starting material for
ZPL389. The typical impurities detected in commercial
batches of starting material 1 are shown in Figure 2.
Figure 1. Structure of ZPL389 (free base).
The most abundant impurity was the 4,5,6-trichloropyr-
imidine-2-amine (14, Figure 2), which was detected at levels
up to 1.80% A [high-performance liquid chromatography
(HPLC)]. All other impurities were detected in significantly
lower amounts. Specifications were set such that after
downstream depletion, no relevant levels of those impurities
and of any related byproducts were present in the final DS
(vide infra).
Starting Material 2. The use of starting material 2
(Scheme 1) was established in the early-phase manufacturing
route. This material, however, posed severe purity issues. The
typical impurities detected in the starting material sourced for
the late-phase campaigns are depicted in Figure 3.
3893787, is a potent, highly selective histamine H4-receptor
antagonist, which was developed as a potential oral drug
candidate for the treatment of moderate to severe atopic
dermatitis.1 The drug substance (DS) was developed in the
form of a tartrate salt dihydrate. This paper aims at describing
the efforts to the development of the process up to the last
chemical step of ZPL389. The final DS crystallization is not in
the scope of this article.
For early-phase manufacturing campaigns, an efficient route
of synthesis from readily commercially available 4,6-dichlor-
opyrimidine-2-amine (1) was developed externally. This route
yielded 7 in a moderate overall yield (54−59%) and very good
purity (Scheme 1). Except for the first manufacturing
campaign, 7 was recrystallized to control the polymorphic
form and PSD of the DS.
Regardless of its source, some batches of cyclopropylme-
thanamine 2 contained >2% A [gas chromatography (GC)] of
1-butanamine (19). This impurity was difficult to purge, led to
persistent byproducts, and, therefore, required strict control. At
Despite the reasonable efficiency, we believed that some
aspects of the manufacturing process required improvement.
The final synthesis route is depicted in Scheme 2.
Received: February 27, 2021
In the early-phase process, some steps ran diluted and the
manufacturing was solvent intensive (see Table 8). Further-
more, certain steps made use of toxic solvents, such as
dimethylformamide (DMF)2a and 1,2-dimethoxyethane
(DME),2b and were very time-consuming. Finally, the accurate
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 1. Early-Phase Route of the Synthesis of ZPL389
the early stages of process development, it was found that the
crystallization of benzoate 2a (Schemes 1 and 2) would allow
the partial purge of 19. Therefore, 2a was used instead of 2 in
the synthesis of intermediate 3 (Schemes 1 and 2).
Although this method showed some success and provided 2a
in an acceptable yield (90−91%), 19 could not be sufficiently
removed by crystallization.
Furthermore, the quality of 2a was strongly dependent on
the erratic levels of 19 in commercially available batches of 2,
which would have led to unacceptably large variability in the
quality of the DS. Contract manufacturing of amine 2 allowed
the tight control of 19 to <0.1% A. However, some batches of
2 stored in the laboratory at room temperature showed
degradation over time. In particular, an unknown impurity,
subsequently identified as imine 20 (Figure 3), was shown to
grow significantly over time.
Table 1 shows the levels of this impurity in four laboratory
batches of 2 obtained from different suppliers, as a function of
storage days in the laboratory at room temperature.
Samples 1 and 2 were from the same batch, and were
received at the same time, packaged in two separate plastic
bottles. Both samples were analyzed on arrival and showed
0.04% A (GC) of impurity 20. Samples 2−4 were stored,
sealed, and sampled only for analysis, as listed in Table 1. The
container of sample 1 was opened several times and after 243
days from delivery 0.72% A (GC) of 20 was detected. Samples
1 and 2 had 1.14% w/w water (Karl Fischer), whereas samples
3 and 4 had a low water content (<0.2% w/w, Karl Fischer).
The oxidation of amines to imines under aerobic conditions
has been recently reviewed.3 Notably, it has also been reported
that this reaction can occur without catalysts in the presence of
molecular oxygen, and that, under certain conditions, this
reaction can be strongly accelerated by water.4
B
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 2. Late-Phase Route of the Synthesis of ZPL389
Figure 2. Impurities found in starting material 1.
Figure 3. Impurities found in starting material 2.
After 243 days, molecular sieves were added to sample 1, in
order to reduce the unusually high water content (1.14% w/w
by Karl Fischer). Following this, more samples were taken
from the container, thereby repeatedly exposing the substance
to air. A steep increase of the rate of the oxidation was
observed and impurity 20 increased to 3.73% A (GC) after 295
days (Table 1). By comparison, in sample 2, which had been
kept tightly closed all the time and was neither treated with
C
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
crystallization of benzoate 2a, which was initially developed
to specifically control impurity 19 (see discussion above).
Contrary to the amine, which is a liquid with a boiling point
of 83−85 °C,5 2a is a crystalline, nonhygroscopic solid, which
can be easily stored and handled. Unlike the parent amine, this
salt was stable for several months under the uncontrolled
conditions of humidity and air in the laboratory.
However, given the high cost of 2, we felt that the yield of 2a
should be increased. By changing the ratio of crystallization
solvents from the initial tetrahydrofuran (THF)/methylcyclo-
hexane = 4.5:4.0 w/w to THF/methylcyclohexane = 1.8:2.7 w/
w, 306 kg of 2a could be manufactured in 98% yield versus
90−91% under the former process conditions.
The resulting crystallization process had a greatly improved
molar yield, volume output, and process mass intensity (PMI)
(see Table 8). Furthermore, this crystallization allowed the
removal of impurity 20 by a factor of over 100-fold. This way,
levels of imine up to almost 7% A (GC) could be significantly
depleted from the isolated benzoate 2a (Table 3). Purge factor
calculations6 showed that even at this level, aldehyde 21 would
be sufficiently depleted in downstream processing.
Table 1. Levels of Impurity 20 in Amine 2 in Laboratory
Samples Stored at Room Temperature
20 (GC, % A)
storage days in lab
sample 1
sample 2
sample 3
sample 4
0
0.04
0.04
0.07
0.01
97
243
295
not tested
0.72
3.73
not tested
not tested
0.58
0.28
not tested
not tested
0.11
not tested
not tested
a
a
Molecular sieves added to reduce water content.
molecular sieves nor sampled, the impurity only increased to
0.58% A (GC) in the same time. Similarly, samples 3 and 4
showed lower rates of oxidation. These had significantly lower
levels of water (<0.2% w/w) and had been kept tightly closed
and not sampled.
The steep increase of 20 in sample 1, following the addition
of molecular sieves could have also been caused by the reaction
of 2 with aldehyde 21, if present in the samples (Scheme 3).
However, neither 21, nor any other unknown impurity was
detected by GC or by 1H NMR in any of the analyzed samples,
including samples 1 and 2.
Table 3. Purging of Impurities in the Spiked Sample of
Starting Material 2 upon Crystallization
Scheme 3. Equilibrium Hydrolysis of Imine 20
purging
factor in step
impurity impurity level in 2
observed level in 2a
2a
19
0.25% w/w
[0.24% A (GC)]
0.26% w/w [0.28% A
(HPLC, after
derivatization)]
0.96 (based
on % w/w)
20
5.04% w/w
<0.05% w/w (HPLC, after >100
derivatization)
Additionally, spiking sample 2 (1.14% w/w water) with
cyclopropanecarbaldehyde (21) resulted in the rapid and
quantitative conversion of this impurity into imine 20 within
minutes, while 21 was barely detectable (Table 2). Therefore,
the increase of impurity 20 in sample 1 could be ascribed to
repeated exposure to air and is unlikely related to the water
content of the samples.
[6.92% A (GC)]
Starting Material 11. Boc-protected (R)-1-benzyl-3-
aminopyrrolidine (11, Scheme 2) was chosen as the GMP
starting material for the late-phase manufacturing process of
ZPL389. In the initial route (Scheme 1), the more advanced
intermediate 4 was used as the regulatory GMP starting
material and manufactured under non-GMP conditions.
However, due to the complexity of intermediate 4, it is
decided to include previous manufacturing steps in the GMP
sequence, as shown in Scheme 2.
Compound 4 was initially manufactured under non-GMP
conditions from raw material 8 in ca. 82% overall yield, via a
sequence of Cbz-protection, N-methylation, and hydro-
genolytic Cbz-cleavage. Because 11 is the synthetic precursor
of 8, it was envisaged that 11 could be more effectively used as
a starting material for the manufacturing of 4 (Scheme 4).
Additionally, owing to the presence of a benzyl group, 11 could
be characterized by HPLC-UV better than 8.
Table 2. GC Analysis of Amine 2 Spiked with
Cyclopropylcarbaldehyde (21)
GC (% A)
others
entry
1
sample type
2
21
20
19
(sum)
a
unspiked sample 2
(295 days)
99.24
n.d.
0.58
0.04
0.14
2
spike with 21 after
89.32
0.008
10.48
0.05
0.14
b
5 min
a
b
Not detected. 5 μL of 21 to 100 mg of 2.
Although in silico evaluation for mutagenicity did not return
alerts for 20, cyclopropanecarbaldehyde 21 was found to be
mutagenic in an Ames test. Carry-over of imine 20 could still
result in the formation of impurity 21 under aqueous
conditions; particularly, if 20 formed in significant amounts
upon storage.
Carefully chosen packaging and storage conditions of amine
2 (cold, inert conditions), along with short storage times, were
anticipated to prevent significant degradation of the amine.
However, the removal of imine 20 along with its derivative,
aldehyde 21, remained of paramount importance, in particular
to control genotoxic 21. This was achieved by the
Compound 11 was sourced in very high purity and against
tight specifications to win approval of the regulatory agencies.
The corresponding enantiomeric impurity was limited to
≤0.10% A (HPLC) and no erosion of the enantiomeric purity
was observed throughout the synthesis.
Development of the Late-Phase Manufacturing
Process of Intermediate 4. The synthesis of 4 from 11
(Scheme 2) was initially developed externally by a telescoped
N-methylation/hydrogenation sequence, in overall 84% yield.
However, poor process robustness, use of DMF and genotoxic
dimethyl sulfate (DMS)7 in the first step, lack of isolation of
intermediate 12manufactured as a solution of its free base
D
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 4. Synthesis of 4 in Early- and Late-Phase Processes
Scheme 5. Formation of Impurities 23 and 24
and poor control over the fate of process impurities prompted
us to revise this synthetic procedure.
acid and render the elimination of t-butoxide reversible. The
addition of t-butanol or t-butoxide to isocyanates is well
documented.10
This hypothesis was corroborated experimentally by stirring
overnight a mixture of 11 and sodium t-butoxide in DMF at 25
°C. Under those conditions, the clean formation of 22% A
(HPLC) of 23, along with unreacted 11 was observed. Heating
that mixture to 50 °C caused further increase of 23 to >35% A
(HPLC).
In the original protocol, 1.7 equiv of sodium t-butoxide and
DMS were added to a solution of 11 in DMF. After stirring,
further additions of both reagents were required at both the
laboratory and plant scale for reaction completion. In a
laboratory experiment, a total of 2.0 equiv of sodium t-butoxide
and 1.9 equiv of DMS were necessary to consume 11 to 0.3%
A (HPLC).
When the whole amount of sodium t-butoxide (2.0 equiv)
was added to the reaction mixture from the beginning and 1.9
equiv of DMS were slowly added in 4 h at −5 °C, ca. 12% A
(HPLC) of residual 11 was still detected. Heating up the
mixture to 25 °C had no impact on conversion. Further
addition of only DMS caused the formation of ca. 14% A
(HPLC) of two diastereoisomeric quaternary ammonium
byproducts (25, Figure 4), while the starting material was
not consumed further.
While attempts at using dimethyl carbonate in place of DMS
were unsuccessful, we sought to avoid the use of DMF. DMF is
REACH-listed as a substance of very high concern (SVHC)
due to reprotoxicity.2a Additionally, the use of strongly alkaline
conditions by t-butoxide could promote the slow formation of
dimethylamine and carbon monoxide by the decomposition of
DMF. Such decomposition is reported to be fast at elevated
temperatures.8
Urea derivatives 23 and 24 (Scheme 5) were detected at ca.
1.4% A (HPLC) in reaction mixtures under the original
methylation conditions in DMF. This provided indirect
evidence for the formation of dimethylamine in the reaction
mixture under those conditions. The deprotonation of 11 and
slow elimination of t-butoxide would lead to isocyanate 22.
Dimethylamine from DMF decomposition could react with 22
and give rise to 23. This in turn would be (partially)
methylated to 24 under the reaction conditions.
Researchers at Pfizer previously described the hydrolysis of
Boc-protected aliphatic and aromatic amines in THF or 2-
methyltetrahydrofuran (MeTHF) in the presence of excess
sodium t-butoxide and water. In the analogy to our hypothesis,
the reaction is thought to occur through an intermediate
isocyanate.9 The tight control of water in our system would
prevent further hydrolysis of the isocyanate to the carbamic
E
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Experiments were designed to assess the effect of molar
ratios, dilution, and temperature on conversion and product
distribution. In all experiments, starting material 11 was
suspended or dissolved in MeTHF, the mixture was cooled to
the desired temperature and KOtBu was added to the
suspension or solution, followed by DMS.
In the initial experiments (entries 1−2), rapid DMS addition
caused the temperature to rise very quickly to ca. 50 °C, likely
leading to the formation of a large amount of byproducts
immediately upon the charging of the reagent. Additional
amounts of base and DMS were required to drive the reaction
to completion. In an attempt at reducing the amount of
byproducts, the slow addition of only 1.0 equiv DMS in 3 h to
the reaction mass at 0 °C was tested (entry 3). The reaction
was incomplete even after overnight stirring at 0 °C.
Interestingly, ca. 0.25 equiv of TBME were detected in that
mixture by GC-headspace (GC-HS), showing DMS con-
sumption by the reaction with t-butoxide even at low
temperatures and by slow addition.
Reducing the base to 1.4 equiv while using 1.5 equiv of DMS
added over 3 h (entry 4) at −5 °C was equally unsuccessful.
Furthermore, upon cooling to −5 °C, a thick suspension
formed.
Increasing the solvent amount to 5 parts w/w helped
mitigating this issue, and greatly improved conversion and
purity were achieved with 2.0 equiv base and 1.7 equiv DMS
(entry 5).
Further increase of the DMS charge to 1.9 equiv (entry 6) or
split charge of DMS in two portions (entry 7) failed to show
improvement and the purity profile was slightly worse. The
byproducts detected in these reactions were the ammonium
salts 25, which were readily removed upon aqueous extraction.
The prolonged exposure of 11 to the base before charging
the DMS (entry 9) was also detrimental to the purity profile
and conversion. In this case, Boc-cleavage impurity 26 formed
in large amounts, as previously discussed (Scheme 6).
Aside from significantly lowering the yield, this byproduct
could not be efficiently removed by subsequent crystallizations.
Therefore, the addition of DMS had to immediately follow the
base treatment.
Figure 4. Quaternary ammonium impurity 25.
Using potassium instead of sodium t-butoxide and running
the reaction at room temperature instead of −5 °C had no
positive effect on the conversion.
It was hypothesized that DMS would react quickly with the
t-butoxide, forming t-butyl methyl ether (TBME), thus
unproductively consuming both reagents and causing an
incomplete reaction. On the other hand, the consumption of
t-butoxide might favor the formation of carbamate 22 (Scheme
6). The detection of the Boc-cleavage byproduct 26 in the
reaction mixtures was believed to support this hypothesis.
Upon aqueous quench, carbamate 22 would hydrolyze to the
corresponding carbamic acid and lose carbon dioxide to form
26 (Scheme 6). On the other hand, due to very low water
content of the reaction mixture, amine 26 should only form
after aqueous quench.9 If formed before, it could have rapidly
reacted with DMS to form N-methylated byproducts 27 and
28, which were not detected.
In order to improve process robustness, the selection of
solvents and bases was screened. The full details of the
screening can be found in the Supporting Information (Table
S1).
From the screening, the choice of potential solvents was
restricted to DMF, THF, and MeTHF. Dioxane, acetone, and
t-butanol gave poorer results, while in the other solvents
screened no product formed. Among the bases tested, only t-
butoxideeither sodium or potassiumgave promising
results. In the presence of weaker inorganic or organic bases,
fast quaternization of the starting material at the endocyclic
nitrogen was observed, while a mixture of byproducts formed
in the presence of LiHMDS.
Further increase of the DMS equivalents (entries 10−12)
led to the significant increase of the byproducts 25 as did the
use of a phase transfer catalyst (entries 11 and 13).
Furthermore, with a larger excess of DMS, the mixture was
unstable upon prolonged stirring and 25 grew steadily over
MeTHF and potassium t-butoxide (KOtBu) were chosen for
the further development of this step. The results of a screening
of conditions are reported in Table 4.
Scheme 6. Formation of Byproduct 26
F
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 4. Screening of Reaction Conditions for Intermediate 12
IPC after complete dosing [HPLC (% A)]
entry
MeTHF (w/w)
KOtBu/DMS (equiv)
T (°C)
dosing time
12
11
others
1
2
3
4
5
6
7
8
4
4
4
4
5
5
5
5
5
5
5
5
1.7/1.7
1.9/1.7
1.9/1.0
1.4/1.5
2.0/1.7
2.0/1.9
2.0/1.9
2.0/1.9
2.0/1.9
2.0/2.1
2.0/2.1
2.0/2.3
0
0
0
2 min
5 min
3 h
3 h
4 h
84.00
88.01
71.76
86.49
97.26
97.05
96.44
93.68
88.94
95.47
94.11
95.43
88.18
90.55
98.43
96.75
95.83
96.64
96.59
95.80
96.27
1.53
6.87
23.21
4.70
0.39
0.48
1.06
0.31
0.17
0.60
0.55
0.34
0.37
0.92
0.49
0.17
0.45
0.36
0.24
0.46
0.64
12.89
5.12
5.03
8.81
2.35
2.47
2.50
6.01
10.89
3.93
5.34
4.23
11.45
8.53
1.08
3.08
3.72
3.13
3.17
3.74
3.02
−5
−5
−5
−5
5
−5
−5
−5
−5
4 h
4 h
a
3.5 h
3.5 h
4 h
4 h
4 h
b
c
9
10
11
12
c
cd
,
13
14
15
16
17
18
19
20
5
5
8
8
2.0/2.1
2.0/2.1
2.0/2.0
2.0/2.0
2.0/2.0
1.8/1.8
1.7/1.7
1.6/1.6
−5
−5
−5
−10
0
−5
−5
−5
4 h
4 h
4 h
4 h
4 h
4 h
4 h
4 h
e
8
8.5
8.5
8.5
a
b
Addition of DMS in two equal portions, each in 2 h and 1 h stirring in between. Mixture of 11, and KOtBu in MeTHF stirred overnight, then
DMS charge. With tetrabutylammonium hydrogen sulfate (0.05 equiv). IPC after overnight stirring. Base added as the 20% THF solution.
c
d
e
time (entries 12−13), while the reaction reproducibility
dropped (entries 11 and 13).
controlled indirectly by methanol specification in 11. Impurity
30 was shown to be the product of Hofmann elimination of
impurities 25 under basic conditions. This was minimized by
running the reaction below 0 °C and further depleted by
crystallization. The quench of the reaction mixture with
aqueous ammonium hydroxide served to destroy residual
DMS.11
Finally, the reaction improved by using 20% w/w KOtBu in
THF (entry 14). A clear solution formed before DMS
addition. However, this protocol was abandoned due to the
substantial increase of the reaction volumes and cumbersome
work-up.
Pivotal to ensuring good DS quality was the removal of all
identified byproducts and of the residual starting material.
While the impurities 25 were readily removed by extraction, all
other impurities required purging from the reaction stream, to
avoid related byproducts in the DS.
Building on this evidence, the solvent amount was increased
further (entries 15−20). Under these conditions, the amounts
of base and DMS were reduced to 1.8 equiv (entry 18) without
significant deterioration of the purity. At the same time, the
residual starting material could be more effectively controlled.
Further lowering the excess of reagents (entries 19−20) caused
residual starting material 11 to be slightly larger.
The control of residual 11 at this stage was pivotal. This
compound leads to related impurities being incompletely
purged further downstream, and therefore, optimal conversion
must be achieved at this stage.
During development, two further impurities were identified
(29 and 30, Figure 5). The first was related to residual
methanol in 11 and formed in analogy to 24 (Scheme 5). This
was confirmed experimentally by spiking methanol to the
reaction mixture before KOtBu charging. Impurity 29 was
The free base of 12 did not crystallize under any of the
investigated conditions. Attempts at crystallizing 12 as
benzoate, in analogy to the following intermediate 4, equally
failed. A small screening of salt forming agents were conducted.
Strong mineral acids such as HCl gave crystalline solids, but
unacceptable Boc-cleavage was also observed. On the contrary,
maleic acid allowed isolation in an excellent yield and optimal
removal of all known process impurities, including residual 11.
The crystallization was performed by charging of a MeTHF-
solution of maleic acid to a solution of the free base in
MeTHF/2-propanol. Dissolving the maleic acid in 2-propanol
caused the rapid formation of i-propyl esters, even at 30 °C.
On the contrary, this was not observed by adding the acid
dissolved in MeTHF to the base. Under the optimized process
conditions, over 1000 kg of 12 could be manufactured in two
batches in >92% yield and with a purity of 99.8% A (HPLC).
The hydrogenation process to 4 as initially developed did
not require any major adjustment. However, an alkaline
extraction in MeTHF to break the maleate was additionally
included in the original procedure. This process allowed the
manufacturing of >740 kg of 4 in 93.3% yield from 12 and a
purity of >99.90% A (HPLC). Notably, the crystallization of 4
as benzoate purged residual unreacted 12.
Figure 5. Structures of impurities 29 and 30.
G
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
In this process, toxic DMF2a was replaced with MeTHF,
thereby avoiding by-products. The process robustness was
greatly improved: the overall amount of reagents was reduced
and the reaction no longer required further additions of base
and DMS. The overall yield was comparable to the telescoped
process (86 vs 84%). Finally, the additional crystallization of 12
accounted for superior control of the process impurities and
enhanced knowledge of their fate. Notably, the quality-critical
starting material 11 could be purged by crystallization.
Development of the Late-Phase Manufacturing
Process of Intermediate 3. The original manufacturing
process of 3 consisted of the reaction between intermediates
2a and 1 in a water/acetonitrile system, in the presence of
potassium carbonate as the base (Scheme 1). The product was
crystallized from TBME/heptane and dissolved in DME
without drying. This solution was then telescoped into the
next step. Based on the assay of the DME solution,
intermediate 3 was obtained in calculated yields of 81−89%.
Despite the high yield and good quality, we decided to
address some shortcomings of the original process. These
included the use of acetonitrile as the reaction solvent and the
large volumes of TBME (18.5 kg/kg of product) and water
(15.1 kg/kg of product) required for the reaction and its work-
up, with a consequently large waste stream. Additionally, a
steep exothermal event in the differential scanning calorimetry
of intermediate 3 with onset >245 °C and ΔH = −750 kJ/kg
raised concerns of strongly energetic spontaneous decom-
position. Therefore, this intermediate was initially neither dried
nor released. However, a negative falling hammer test ruled out
shock sensitivity, while no evidence for deflagration was found
in a closed pressure vessel test. Therefore, it was deemed safe
to dry and release 3.
We sought to replace acetonitrile with a cheaper and
possibly water-immiscible solvent in order to reduce solvent
volumes and waste generation. Furthermore, we anticipated
that the product could be isolated in higher yields.
A screening of reaction conditions (solvents and bases) was
carried out. A summary of the results is reported in the
Supporting Information (Table S2).
In the presence of N-methylmorpholine, all reactions
performed poorly and had the worst purity profile. In these
experiments, large amounts of two impurities were detected by
liquid chromatography−mass spectrometry (LC−MS). One of
them was confirmed to be benzamide 31, while the other was
putatively assigned the structure 32 based on LC−MS data
(Figure 6); however, this was not confirmed.
conversion after 24 h, despite the high temperature (90 °C).
Furthermore, the reaction work-up was cumbersome due to
poor phase separation and required a large amount of TBME
for extraction. Failure to reach completion was thought to be
caused by the temperature being above the boiling point of
amine 2 (83−85 °C4). Reactions in 1-butanol did not show
improvement.
In ethereal solvents, such as THF/water and MeTHF/water,
moderate conversion and good purity could be achieved.
In ethyl acetate at 80 °C and in the presence of solid
potassium carbonate, high conversion and purity were
obtained.
Notably, entering experimental solubility data of 3 into the
DynoChem 5 utility “early-phase solvent selection: solubility
prediction” (Scale-up Systems Ltd.; Dublin, Ireland) indicated
that mixtures of acetic acid esters, such as ethyl and isopropyl
acetate, and alkanes, for example, n-heptane, could be viable
crystallization solvents for the isolation of 3. The use of acetic
acid esters as the reaction solvent presented the opportunity of
running reaction, work-up, and crystallization in the same
solvent. This would in turn simplify the process and reduce the
overall solvent burden.
The reaction mixture in ethyl acetate was heterogeneous due
to solid potassium carbonate. We anticipated that upon scale-
up, this might pose issues such as stirrability of the mixture and
adequate suspension of the solids, as well as the dependence of
the reaction on the particle size of the carbonate. Taken
together, these might lead to poor process robustness.
With this in mind, the experiments summarized in Table 5
were run. In those experiments, 1.1 equiv of 2a was used and
all experiments were run on a multigram scale, at the reflux and
with mechanical stirring.
All reactions had comparable purity profiles. Only one
impurity (36, Figure 7) formed with ca. 0.2% A; therefore, only
the ratio of the product and starting material is reported.
Strikingly, already upon scale-up to gram scale in the
laboratory, the reaction in ethyl acetate with potassium
carbonate (Table 5, entry 1) did not reproduce the outcome
of the small-scale experiment, and conversion was only ca. 92%
after 4 h. Complete conversion was not achieved even after
prolonged stirring. Because it had been anticipated that this
might be ascribed to the use of a solid insoluble base, other
options were tested.
With soluble organic bases (entries 2 and 3), very poor
conversion was achieved after 4 h at the reflux, while no new
impurities were detected. Even after 16 h, conversion was less
than 85% (data not shown). When the reaction was carried out
in the presence of potassium carbonate in ethyl acetate/water
(entry 4), a clear biphasic mixture was obtained; however, the
reaction temperature reached only ca. 70 °C, the boiling point
of the water/ethyl acetate azeotrope. The reaction was
incomplete after 18 h. Switching to isopropyl acetate/water
(entry 5) allowed reaching a slightly higher temperature (78−
80 °C). This reaction was significantly faster and complete
after 6 h. Anticipating similar reproducibility issues as with
ethyl acetate, anhydrous isopropyl acetate was not evaluated.
Despite the poor solubility of 1, this dissolved within 2 h and a
clear biphasic mixture was obtained.
Figure 6. Structures of byproducts 31, 33−35, and putative
byproduct 32.
Potassium carbonate and N-ethyl-N,N-diisopropylamine
(EDIPA) were equally efficient in the synthesis reaction to 3.
In alcoholic solvents (methanol, ethanol, and 1-butanol), the
respective alkyl ethers 33−35 (Figure 6) were detected.
Surprisingly, the analogous byproducts from 2-propanol and 1-
propanol were not detected. In 1-propanol/water, the reaction
proceeded quickly within 4 h, but failed to reach full
These conditions were scaled up further and showed
excellent reproducibility. Switching from potassium to sodium
carbonate gave the same outcome. The work-up consisted of
simple aqueous extractions of the reaction mixture, followed by
distillation, crystallization from isopropyl acetate/n-heptane,
H
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 5. Screening of Reaction Conditions for Intermediate 3
HPLC (% A)
entry
solvent
base (equiv)
T (°C)
time (h)
3
1
1
2
3
4
5
ethyl acetate
ethyl acetate
ethyl acetate
ethyl acetate/water (2/1 v/v)
isopropyl acetate/water (3/2 v/v)
K2CO3 (2.5)
triethylamine (2.5)
EDIPA (2.5)
K2CO3 (2.5)
K2CO3 (2.5)
80
80
80
70
80
4
4
4
4
4
91.97
76.35
68.71
91.01
96.20
8.03
23.65
31.29
8.99
3.80
Figure 7. Typical process impurities in 3.
Scheme 7. Reaction of 3 with 4 or 41
and drying. The crystallization was efficiently designed with the
DynoChem “crystallization toolbox” based on few experimen-
tal data.
Plant scale-up afforded ca. 245 kg of intermediate 3 in three
batches, with an average 95.2% yield and excellent purity. The
only impurity detected in the plant batches was 36 (Figure 7),
at 0.2% A in all batches. Although having a positive in silico
alert for mutagenicity, this impurity was negative in an Ames
test.12
Except for 36, other typical process impurities detected in
intermediate 3 depended on the quality of starting materials 1
and 2 and are depicted in Figure 7. Impurity 37 originated
from the reaction of 1 with the butanamine (19, Figure 3) in
2a and was controlled by the specification of 19 in incoming 2,
although 37 was depleted by ca. 1.8-fold in the crystallization
of 3. Impurities 15 and 16 were carried over from
dichloropyrimidine 1. These were unreactive under process
conditions and were removed by aqueous extraction. 38 was
the main impurity expected based on the typical quality of 1. It
originated from 4,5,6-trichloropyrimidine-2-amine (14, Figure
2) and was depleted by a factor larger than 18-fold upon the
isolation of 3. Byproducts 39 and 40 derived, respectively,
from impurities 17 and 18 in starting material 1 (Figure 2) and
were depleted by a factor 2.4 and 1.3, respectively, upon
isolation. In the absence of 17 and 18, the corresponding 39
and 40 were not detected. Effective rejection to <LOQ (0.05%
A, HPLC) was achieved in the following step.
Development of the Late-Phase Manufacturing
Process of Intermediate 5. In the original process,
intermediate 5 was manufactured by reacting 3 and 4 in 1,2-
DME in the presence of EDIPA as the base at 80 °C (Scheme
1). A large excess of 4 (1.3 equiv) and long reaction times (>5
days) were required to reach adequate conversion. The
product was extracted as free base in TBME, followed by the
addition of fresh benzoic acid, whereupon benzoate 5
crystallized. The product had high purity and the yield was
91% (overall 74−81% from 3).
This process also suffered severe limitations, including the
use of reprotoxic and REACH-listed SVHC DME,2b long
reaction time, use of large excess of expensive 4, use of large
amounts of solvents and significant waste generation. These
issues were addressed by process development.
Similar nucleophilic aromatic substitutions are typically
performed in high-boiling polar solvents at elevated temper-
atures (80−160 °C) and in the presence of bases such as
fluoride, carbonate, or EDIPA.13 Reports of such reactions
I
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
proceeding under microwave irradiation at 160 °C were found
too.13b,14
Based on these precedents, a screening of solvents and bases
was carried out. Due to the limited availability of intermediate
4 during development, surrogate compound 41 was mostly
used to test reaction conditions (Scheme 7). Details are given
in the Supporting Information (Table S3).
stronger formic acid (entry 10), conversion dropped.15 In both
reactions, significantly more decomposition was detected.
In analogy to benzoic acid, the reaction was also tested in
the presence of Lewis acids, although without major improve-
ments. The details are available as Supporting Information
(Table S4).
To investigate further the impact of pH on the reaction, the
reaction was screened in the presence of phosphate salts. The
results of the screening of the reaction between 3 and 4
(benzoate salt) in the presence of phosphates are shown in
Table 7.
Simple heating up of the reagents in the chosen solvent
resulted in slow conversion and/or major impurities formation.
Above 100 °C, large amounts of byproducts were formed. In
none of the conditions tested complete conversion was
achieved, even upon long stirring (24−32 h). Inorganic bases
(K2CO3 and Na2CO3) gave lower conversions than soluble,
high-boiling tributylamine. Without base, the reaction stalled at
71−82% conversion with significant product decomposition.
The reactions in MeTHF at 80 °C were slower (46−83%
conversion after 39 h), but the purity profile was very good,
likely related to the lower reaction temperature. With sodium
carbonate the conversion was lower (46% after 39 h) than with
EDIPA. This was probably due to the lack of solubility of the
carbonate.
In order to assess the effect of the benzoate salt on
conversion and purity profile, experiments with surrogate
starting material 41 (1.15 equiv) and various amounts of added
benzoic acid were conducted in 1-methoxy-2-propanol
(propylene glycol monomethyl ether) at 125 °C (Table 6).
Reagents and solvent were stirred to obtain a clear solution, a
base was added where applicable (entries 2−3), and the
mixture was heated to 125 °C.
Initial experiments were carried out in MeTHF at 80 °C, as
these conditions had proven beneficial to product purity in the
presence of bases (see discussion above). Generally, although
conversions up to 89% A (entry 1) could be achieved in 39 h
and no impurity above 0.05% A (HPLC) was detected, the
reactions suffered from very poor reproducibility (entries 2−3
and entries 3−4) and conversions were erratic (entries 1−6).
This was explained by the lack of solubility and the
uncontrollable particle size of the inorganic additive. The
addition of water to that mixture would lower the boiling point
to ca. 73 °C (boiling point of the azeotrope MeTHF/water),16
which would have affected negatively the reaction rate. For this
reason, further experiments were conducted in toluene/water
(entries 7−13). This mixture forms an azeotrope with boiling
point ca. 84 °C, as found experimentally. The reaction in the
presence of 1.2 equiv of 4 and 2.5 equiv of K2HPO4 at 84 °C
was very fast and well reproducible (entries 7−8). The pH of
the aqueous phase was ca. 7 throughout the experiment. The
use of 2.5 equiv KH2PO4 or K3PO4 caused the conversion to
be significantly lower (entries 9−10), while a pH of 4 and 10,
respectively, was measured at the end of the experiment. With
1.0−1.5 equiv of K2HPO4 (entries 11−13), conversion was
much lower and the reactions did not reach completion.
Furthermore, when lower amounts of K2HPO4 were used, the
pH dropped in the course of the reaction. Finally, in xylenes/
water (entry 14), the reaction could be carried out at ca. 98 °C,
the boiling point of that mixture. The reaction was faster (ca.
99% A conversion in 17 h) and pleasingly, no product
decomposition was observed. After 24−26 h, starting material
3 was completely consumed.
Table 6. Effect of Benzoic Acid on the Reaction of 3 and 41
HPLC (% A)
benzoic acid
(equiv)
base
(equiv)
time
(h)
entry
3
5′
others
1
2
24
32
20.79
19.88
75.92
76.14
3.29
3.98
1.15
1.15
EDIPA
(2.5)
3
EDIPA
(1.0)
32
43.34
52.81
3.85
4
5
6
7
8
9
10
1.15
0.50
2.0
24
24
24
24
19
24
24
37.99
8.39
8.8
2.23
<0.05
13.65
59.51
88.14
83.57
89.78
64.47
74.90
39.26
2.5
3.47
7.63
7.99
35.53
11.45
25.66
Intermediate 5 was isolated as the benzoate salt because its
free base did not crystallize. Under the newly established
process conditions, benzoate 5 crystallized directly in the
course of the reaction, and only required isolation by
centrifugation and washing; thus, avoiding tedious work-up
procedures and the use of additional benzoic acid, as in the
original protocol. The yield (corrected by the assay after salt
breaking) was ca. 99 versus 91% of the original procedure and
the crystallization removed all residual process impurities with
very high purging efficiency.
3.0
3.0
a
b
1.15
c
1.15
35.08
a
b
c
5 g scale. Acetic acid instead of benzoic acid. Formic acid instead
of benzoic acid.
Unfortunately, the isolated product was not homogeneous
because of varying amounts of phosphates. This affected the
assay measurements by the titration of 5 and the control of the
stoichiometry in the following step. It was argued that the main
contaminant would be KH2PO4, formed by the reaction of
K2HPO4 and HCl generated in the reaction, this, however,
could not be confirmed. The calculated concentration of
KH2PO4 in the aqueous phase at the end of the reaction was
above its solubility. Slurrying the isolated product in water at
30 °C efficiently removed the inorganics. However, the
avoidance of an additional slurrying step after product isolation
was desirable to minimize plant operations.
The use of benzoic acid without a base increased the
reaction rate, although at the expenses of purity. Surprisingly,
the addition of 0.5 equiv of benzoic acid gave fast and
moderately clean reaction; however, under these conditions,
the reaction was incomplete after 24 h. Further increase of the
benzoic acid to 3.0 equiv resulted in the higher consumption of
starting material 3 (entry 7); however, the product
decomposed mostly by Boc-cleavage. Repeating this experi-
ment on a 5 g-scale (entry 8) caused even larger
decomposition and significant charring, although 3 was
consumed already after 19 h. Interestingly, with slightly weaker
acetic acid (entry 9), the reaction was faster, while with
J
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 7. Screening of Phosphates in the Reaction of 3 and 4
HPLC (% A)
entry
4 (equiv)
solvent
additive (equiv)
T (°C)
time (h)
pH
3
5
others
1
2
3
4
5
6
7
8
1.2
1.2
1.2
1.2
1.2
1.05
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
MeTHF
MeTHF
MeTHF
MeTHF
MeTHF
MeTHF
toluene/water (3/2 v/v)
toluene/water (3/2 v/v)
toluene/water (3/2 v/v)
toluene/water (3/2 v/v)
toluene/water (3/2 v/v)
toluene/water (3/2 v/v)
toluene/water (3/2 v/v)
xylenes/water (4/2 v/v)
K2HPO4 (2.0)
K2HPO4 (2.5)
K2HPO4 (2.5)
K2HPO4 (1.5)
K2HPO4 (1.5)
K2HPO4 (2.5)
K2HPO4 (2.5)
K2HPO4 (2.5)
KH2PO4 (2.5)
K3PO4 (2.5)
80
80
80
80
80
80
84
84
84
84
84
84
84
98
39
39
32
20
20
39
17
24
24
24
22
22
21
17
10.83
40.81
26.00
14.89
33.04
15.22
1.13
89.17
59.19
74.00
85.11
66.96
84.78
98.87
98.09
79.40
71.23
91.27
98.96
97.54
98.92
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
7
7
4
10
7
7
7
1.91
9
20.59
28.77
8.73
1.04
2.46
10
11
12
13
14
a
K2HPO4 (1.0)
K2HPO4 (1.5)
K2HPO4 (1.9)
K2HPO4 (1.9)
b
1.08
a
b
pH dropped to 5.5 at the end of the experiment. pH dropped to 6.5 at the end of the experiment.
To address this issue, we took advantage of the uniqueness
material had a purity of >99.8% A (HPLC) and phosphate
content in all batches ranged between 0.06 and 0.12% w/w.
Sulfated ashes were found to be between 0.16 and 0.84% w/w
and correlated with the phosphate content, showing that
residual phosphate was essentially the only inorganic
contaminant.
of the reaction mixture. At the end of the reaction, within
minutes from stopping the stirrer, three distinct layers formed:
a lower cloudy aqueous phase, a middle layer consisting of a
suspension of the solid product in water/xylenes, and an upper
clear brownish organic layer (Figure 8).
In summary, a novel process for the manufacturing of 5 was
developed and demonstrated on scale. This allowed much
shorter reaction times of 24−26 h versus >5 days, the
replacement of the reprotoxic, REACH-listed solvent DME2b
by xylenes, use of a phosphate base, slight reduction of the
molar excess of 4 from 1.3 to 1.2 equiv, superior conversion
and yield, and significant lowering of the step’s PMI (see Table
8).
The typical process impurities found in 5 are shown in
Figure 9. Impurity 36 was carried over unreacted from 3.
Impurities 42−45 were formed by the reaction of intermediate
4 with impurities potentially present in 3 (see Figure 7) and
were dependent on the quality of the sourced starting materials
1 and 2. Impurity 42 was not depleted by crystallization but
was controlled by the specifications of 1-butanamine in starting
material 2 (see discussion above). All other impurities, if
present, could be very effectively purged by >10- to 20-fold by
crystallization, resulting in typical levels of <0.05% A (HPLC)
in the isolated product.
Development of the Late-Phase Manufacturing
Process of Intermediate 6. Aside from replacing acetyl
chloride/ethanol with gaseous HCl, the original manufacturing
process of 6 was believed to be sufficiently robust for scale-up.
However, during laboratory activities, it appeared evident that
also in this step unresolved issues needed addressing.
In the original protocol, the benzoate salt of 5 was broken
using aqueous 10% K2CO3 and the free base was extracted in
MeTHF. The solvent was then swapped to ethanol, whereby
water content could also be controlled. Replacing the
potassium carbonate solution with more concentrated aqueous
NaOH reduced the total volume of the extraction and avoided
possible degassing from K2CO3. Replacing MeTHF by TBME
and halving the solvent amount simplified the subsequent swap
distillation to ethanol and reduced solvent usage by ca. 70%.
The crystallization of 6 after Boc-cleavage in ethanolic HCl
at 60 °C was uncontrolled, regardless of the source of HCl,
which was of concern.
Figure 8. Reaction mixture of 5 after the complete reaction.
After reaction completion, 0.7 equiv KOH were charged to
the stirred mixture, in order to convert part of KH2PO4 back to
the more soluble K2HPO4. The pH before and after addition
was ca. 7, which avoided product loss in the organic phase.
Most of the lower aqueous phase was removed. Fresh water
was charged, and the 3-phase mixture was stirred, centrifuged,
and washed with xylenes and water, to afford pure 5 as a white
solid without loss of yield and very low inorganic content.
This process was scaled-up and over 630 kg of 5 were
manufactured in five batches, in 98.5% average yield. The
K
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 9. Typical process impurities in 5.
Figure 10. Byproducts in the synthesis of 6.
Intermediate 6 was crystallized directly from a strongly
supersaturated ethanolic solution after complete reaction and
initiated by seeding. Turbidimetric analysis, however, showed a
sudden increase of the turbidity occurring inconsistently either
after seeding or upon cooling. Microscope imaging showed
small needles ca. 5−10 μm of length that tended to
agglomerate. This resulted in variable filtration times for
typical lab-batches on a 20 g scale, which in some cases
exceeded 1 h. All attempts at controlling this crystallization
were unsuccessful.
The solubility of the product in various alcohols was
measured with a Crystal 16 and fitted to a van’t Hoff equation
to provide solubility curves. Compound 6 showed a moderate
solubility in ethanol of about 50 mg/mL at 65 °C. In contrast,
solubility in methanol was much higher and it significantly
decreased by cooling. It was anticipated that switching from
ethanol to methanol might allow better control of the
crystallization process by lowering supersaturation at the
start of the crystallization, while not being detrimental to the
Boc-cleavage step.
effectively controlled by the distillation of its azeotrope with
methanol.
Optimal conditions for the Boc-cleavage were determined by
kinetic modeling.17 Thus, in the presence of 3.0 equiv HCl, the
rapid cleavage of the Boc-group was achieved at 50 °C in
methanol. From the kinetic model, it was also found that
adding the methanolic solution of the free base of 5 to 12.5%
w/w methanolic HCl at 50 °C in 6 h would ensure the control
of the gas flow rate from the Boc-cleavage on scale. This was
confirmed experimentally. After complete charging of the free
base of 5 to HCl, the reaction was seeded and further stirred
until completion. The product crystallized slowly. Following
temperature cycling, cooling and addition of TBME as the
antisolvent, the product was isolated in a high yield (93.4%)
and purity (>99.95% A, HPLC). The main impurities were
methyl ethers 47, 49, and 51 (Figure 10). The largest was 47,
typically detected at 0.2−0.3% A (HPLC) in the reaction
mixture. 49 and 51 were ca. 0.1% A (HPLC) in typical reaction
mixtures. All impurities were purged upon crystallization to
<0.10% A (HPLC).
Both the Boc-cleavage and the crystallization were sensitive
to residual water. This would cause the formation of hydrolysis
byproducts 46, 48, and 50 (Figure 10) and lower the
crystallization yield.
Water control was initially achieved by azeotropic distillation
with ethanol to <0.3% w/w. At that level, impurities 46, 48,
and 50 were under control and the crystallization yield was
high. Replacing ethanol by methanol did not allow effective
water removal. However, if toluene was used in the extraction
step in place of TBME, optimal control of water levels was
achieved again. Additionally, residual toluene could be very
375 kg of intermediate 6 were manufactured according to
this protocol in three batches, with an average yield of 92% and
a purity >99.9% A (HPLC). Slight yield loss in the first batch
during centrifugation accounted for the lower-than-expected
plant yield. The adjustment of the centrifugation in the second
and third batch allowed product recovery in 95 and 96% yields,
respectively. Compared to the original protocol, volumes,
solvent usage, and waste generation were strongly reduced
(Table 8), and the crystallization was stabilized.
In the plant batches, impurities 47, 49, and 51 amounted to
500−830, <5 and 100−200 ppm, respectively. Of those
L
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
impurities, compound 47 had given a positive in silico alert for
mutagenicity.12 The Ames test, however, returned a negative
result.
Table 8. PMI Calculations
intermediate
original process
new process
2.7
20
2a
12
4
4.5
telescoped
Development of the Late-Phase Manufacturing
Process of Intermediate 13. The crystallization of the
tartrate dihydrate 7 (Scheme 1) was addressed. This was the
intended salt form of ZPL389-DS. In the original process, the
free base of 6 was extracted from aqueous solution into
MeTHF after pH adjustment with NaOH. Solvent switch to
methanol was lengthy and solvent consuming. The desired
tartrate dihydrate was crystallized from methanol/water in 83%
yield. This material was then taken into a recrystallization step,
which served to adjust the particle size distribution of the
material, while controlling the polymorphic form of the DS.
When this procedure was repeated, strong discoloration
occurred during the long distillation. Although no new
impurities >0.10% A were detectable in the HPLC, the
conversion of this material into the tartrate proceeded in a
lower-than-usual yield, crystals could not be adequately grown,
and the isolated material had a faint red color.
The recrystallization of this material to the DS was equally
low yielding and, while the polymorphic form was unaffected,
failure to achieve sufficient crystal growth represented a severe
limitation that compromised the physical properties of the DS
itself.
Thus, it was concluded that prolonged thermal stress of the
free base of 6 could cause significant quality issues.
Alternatives to the extraction work were considered.
Unfortunately, 6 as a free base is poorly soluble in most
water-immiscible organic solvents. It was found that the
extraction could be carried out in 1-butanol. After azeotropic
water distillation, a clear, colorless solution of 6 free base in
high purity was obtained. Prolonged heating of that solution
did not cause any discoloration, nor appearance of impurities
by HPLC. Attempts at crystallizing the tartrate directly from
the 1-butanol solution, however, caused the formation of an
unstirrable, thick suspension. The addition of methanol and
water to the crystallization mixture improved significantly the
stirrability.
After extraction of the free base of 6 in 1-butanol, water was
reduced to ≤5% w/w by distillation, and the concentration of
6 free base was adjusted to 12% w/w by dilution with 1-
butanol. This solution was charged to a warm solution of
tartaric acid in methanol/water in two portions, with seeding
in between. After temperature cycling and cooling to 5 °C, the
product was isolated and dried under reduced pressure. Under
these conditions, a new nonstoichiometric water/methanol
solvate (13, Scheme 2) was isolated in 96.7% yield and
excellent purity. The level of 1-butanol was consistently <500
ppm.
39.6
59.6
27.5
7.0
34.5
39.2
26.8
162.8
total 12 + 4
3
5
total 3 + 5
6
7/13
total
35.3
telescoped
49.5
69.4
49.4
208.1
steps, the PMIs could be reduced by 30−46% (overall ca. 22%)
by streamlining operations and minimizing the usage of
organic solvents and water.
CONCLUSIONS
■
The early-phase manufacturing process to ZPL389 was further
developed, in order to avoid toxic solvents and reagents,
increase the overall yield, and reduce solvent consumption and
waste generation. Toxic solvents were replaced with more
benign options. Enhanced knowledge over process impurities
and their fate resulted in improved process robustness and
quality. The sourcing strategy of starting material 2 was crucial
for the control of a critical process impurity, while the
crystallization of benzoate 2a allowed mitigating stability issues
of 2 and controlling a genotoxic impurity. The accurate control
of process conditions ensured very high DS quality. The
extensive optimization of reaction conditions for the
methylation of 11 led to a robust, scalable, and high yielding
process, as demonstrated on scale. Recognizing the impact of
pH on the formation of 5 allowed a reduction of the reaction
time from >5 days to just about 25 h. As a result each
manufacturing batch could be completed in just 30−35 h
versus >1 week in the early phase process. Selective salt
formation at various stages was pivotal to controlling the fate of
process impurities. In particular, multiple crystallizations of the
DSthrough hydrochloride 6 and tartrate 13proved
essential to achieve high purity and control the polymorphic
form and particle size of the DS. Work-up and isolation
protocols were streamlined and overall yield of 13 from 1
increased from 54-59 to 81%. The final intermediate 13 was
manufactured on scale in very high purity (>99.90% A,
HPLC). PMI metrics show a significant reduction of material
usage in all but one step of the process.
EXPERIMENTAL SECTION
■
General. All materials were obtained from commercial
sources and were used without further purification. All solvents
and process aids were obtained from commercial sources, were
of pilot plant quality, and were used without any further
purification. Solubility data were collected by polythermal
method with a Crystal 16 (Technobis crystallization systems,
Alkmaar, Netherlands). HPLC chromatograms were recorded
on Agilent 1100 or 1290 series instruments. NMR spectra were
recorded on a Bruker 500 MHz UltraShield instrument
equipped with a cryoprobe or on a Bruker Avance III 400
MHz instrument and recorded at room temperature in
deuterated solvents. The spectra are referenced against residual
nondeuterated solvent peak. High-resolution mass spectra
(HRMS) data were recorded on an Agilent 6530 QTOF high-
478 kg of 13 were manufactured in three batches according
to the protocol described above. The average yield was 94.4%,
thus slightly lower than expected. The purity was consistently
>99.90% A (HPLC) and water content of the isolated dry
product ranged between 1 and 2% w/w, while methanol
content was 6.6−7.0% w/w.
This new solvate could be successfully recrystallized in high
yield into the desired form of the DS. Details of the
recrystallization are out of the scope of this paper.
Table 8 shows the calculated PMI for each step of the new
process in comparison with the original protocol. Notably, the
isolation of intermediate 12 introduced in the new process was
done at the expenses of the step’s PMI. However, in all other
M
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
resolution mass spectrometer, coupled with an Agilent 1290
UHPLC instrument. GC-HS data were recorded on a Hewlett
Packard HP6890 GC instrument, equipped with an Agilent
7694 autosampler. All reactions were performed under
nitrogen or argon.
potassium t-butoxide (58.5 g, 521.0 mmol, 1.8 equiv). The
mixture was heated to 40 °C in ca. 30 min and stirred for about
15 min. The resulting turbid solution was cooled to −5 °C. A
solution of (3R)-1-benzyl-3-[(tert-butoxycarbonyl)(methyl)-
amino]pyrrolidine (11, 80.0 g, 289.5 mmol, 1.0 equiv) in
MeTHF (200.0 g) was charged at −5 °C. The resulting
mixture was stirred for ca. 15 min, then a solution of DMS
(65.9 g, 521.0 mmol, 1.8 equiv) in MeTHF (80.0 g) was
charged at −5 °C in 4 h. To the reaction mixture, a solution of
ammonium hydroxide 28% (65.4 g) in water (120 g) was
charged and the mixture was warmed up to 25 °C. The phases
were split and the organic phase was extracted 3 times with
water (160 g). The reaction mixture was concentrated to ca.
125 g under reduced pressure. The solution was diluted with
MeTHF (85.0 g) and 2-propanol (80.0 g) and heated to 48
°C. A solution of maleic acid (33.6 g, 289.5 mmol, 1.0 equiv)
in MeTHF (120.0 g) was prepared, and half of it was charged
to the solution at 50 °C. Seeds were charged. The solution
became turbid and was stirred for 30 min. The remaining
maleic acid solution was added in 1 h. The suspension was
cooled to 0 °C in 4 h and stirred for 2 h. The solids were
collected by pressure filtration and rinsed twice with a pre-
cooled solution of MeTHF (60.0 g) and 2-propanol (20 g).
The solids were dried at 50 °C under reduced pressure. 12 was
obtained as a white solid (109.3 g, 92.9% yield) with purity
Cyclopropylmethanaminum Benzoate (2a). To a 1 L
double-jacketed reactor equipped with an anchor stirrer,
benzoic acid (175.0 g, 1433.0 mmol, 1.02 equiv) and THF
(177.2 g) were charged. The mixture was cooled to 10 °C. To
the resulting suspension a solution of cyclopropylmethanamine
(2, 100.0 g, 1406.1 mmol, 1.0 equiv) in methylcyclohexane
(266.2 g) was charged at 10 °C. At the end of the charging, a
suspension was obtained. The mixture was cooled to 0 °C in
60 min and stirred for 1 h. The solids were collected by
filtration on a pressure filter and dried under reduced pressure
at 35 °C (Caution: the solid sublimates at higher temper-
ature). 2a was obtained as white solid (250.0 g, 92.0% yield)
with purity >99.9% A (HPLC). 1H NMR (500 MHz, DMSO-
d6): δ 0.34−0.24 (m, 2H), 0.51−0.38 (m, 2H), 1.06 (m, 1H),
2.70 (m, 2H), 7.40−7.22 (m, 3H), 7.91 (m, 2H), 8.69 (s, 3H).
13C NMR (101 MHz, DMSO-d6): δ 4.17, 9.46, 44.03, 127.88,
129.48, 129.90, 139.27, 171.01. HRMS analysis was done by
GC−MS with chemical ionization, after derivatization as
trifluoroacetamide. For C6H9ONF3 [M + H]+ m/z: 168.0629;
theoretical, 168.0631.
1
6-Chloro-N4-(cyclopropylmethyl)pyrimidine-2,4-diamine
(3). To a 1 L double-jacketed reactor equipped with an anchor
stirrer were charged cyclopropylmethanaminium benzoate (2a,
64.81 g, 335.4 mmol, 1.10 equiv), sodium carbonate (80.79 g,
762.2 mmol, 2.50 equiv), water (200.0 g), 4,6-dichloropyr-
imidine-2-amine (1, 50.0 g, 304.9 mmol, 1.0 equiv), and
isopropyl acetate (183.0 g). The mixture was heated to 78−80
°C (light reflux) in 30 min with moderate stirring and stirred
for 12 h. The resulting mixture was cooled to 55 °C and
diluted with isopropyl acetate (60 g) and a solution of sodium
carbonate (5.0 g) in water (145 g). The resulting mixture was
stirred at 55 °C for 15 min, and then the phases were split. The
aqueous phase was extracted with isopropyl acetate (100 g) at
55 °C. To the combined organic extracts was charged a
solution of sodium sulfate (3.2 g) in water (96.8 g). The
mixture was cooled to 10 °C and filtered over a cellulose pad.
The filter was rinsed with isopropyl acetate (100 g). The
filtrate was cooled to 10 °C and phases were split. The solvent
was evaporated until the distillate flow ceased, and the residue
was diluted with isopropyl acetate (60 g) and heated to 55 °C.
n-Heptane (100 g) was charged to the clear orange solution at
55 °C. The solution was seeded and stirred for 1 h at 55 °C.
The resulting suspension was cooled to 0 °C in 4 h, and then
n-heptane (210 g) was charged in 1 h. The suspension was
stirred for 1 h. The solids were collected by filtration on a
pressure filter and rinsed with a solution of isopropyl acetate
(10 g) and n-heptane (40 g). The solids were dried at 75 °C
under reduced pressure. 3 was obtained as a yellowish solid
(55.5 g, 91.6% yield) with purity 99.64% A (HPLC). 1H NMR
(500 MHz, DMSO-d6): δ 0.23−0.09 (m, 2H), 0.49−0.35 (m,
2H), 0.96 (m, 1H), 3.10 (s, 2H), 5.76 (s, 1H), 6.39 (s, 2H),
7.18 (s, 1H). 13C NMR (101 MHz, DMSO-d6): δ 3.63, 3.71,
3.76, 11.01, 11.09, 45.02, 93.30, 157.71, 163.34, 164.43.
HRMS [M + H]+ m/z: 199.0739; theoretical, 199.0745.
(3R)-1-Benzyl-3-[(tert-butoxycarbonyl)(methyl)amino]-
pyrrolidin-1-ium (2Z)-3-carboxyprop-2-enoate (12). To a 1 L
double-jacketed reactor equipped with an anchor stirrer were
charged under argon flow anhydrous MeTHF (400.0 g) and
99.88% A (HPLC). H NMR (500 MHz, DMSO-d6): δ 1.38
(s, 9H), 2.06−1.96 (m, 1H), 2.15 (m, 1H), 2.75 (s, 3H), 3.13
(m, 1H), 3.24 (d, J = 9.6 Hz, 1H), 3.45−3.31 (m, 2H), 4.36 (t,
J = 9.6 Hz, 2H), 4.74 (m, 1H), 6.09 (s, 2H), 7.47−7.40 (m,
3H), 7.54−7.48 (m, 2H). 13C NMR (126 MHz, DMSO-d6): δ
25.97, 27.98, 30.14, 52.33, 52.88, 53.23, 57.34, 79.45, 128.89,
129.30, 130.33, 131.64, 135.69, 154.61, 167.39. HRMS [M +
H]+ m/z: 291.2072; theoretical, 291.2067.
(3R)-3-[(tert-Butoxycarbonyl)(methyl)amino]pyrrolidin-1-
ium Benzoate (4). To a 1 L double-jacketed reactor equipped
with an anchor stirrer were charged 12 (80.0 g, 196.8 mmol,
1.0 equiv) and MeTHF (320.0 g). The mixture was stirred at
20 °C. A solution of sodium hydroxide (17.32 g, 433.0 mmol,
2.2 equiv) in water (156.0 g) was slowly charged to the
mixture at 20 °C. The phases were split and the organic
solution was extracted 3 times with water (160 g) to pH ca. 6.
The organic phase was transferred to an hydrogenation vessel
and benzoic acid (24.76 g, 202.7 mmol, 1.03 equiv), 5% Pd on
charcoal (1.6 g) and methanol (120.0 g) were charged. The
hydrogenation was conducted at 50 °C under 4 bar hydrogen
pressure for 15 h. The mixture was cooled to 25 °C and filtered
over a cellulose pad. The filter residue was rinsed twice with
MeTHF (130.0 g). The filtrate was transferred to a 1 L double-
jacketed reactor equipped with an anchor stirrer and heated to
60 °C and distilled under reduced pressure to ca. 290 g. The
solution was cooled to 50 °C and methylcyclohexane (200.0 g)
was charged. The solution was seeded and stirred for 3 h. To
the resulting suspension, methylcyclohexane (200.0 g) was
charged at 50 °C in 1.5 h. The mixture was cooled to 10 °C in
4 h and stirred for 1 h. The solids were collected by pressure
filtration and rinsed twice with methylcyclohexane (125 g).
The solids were dried at 40 °C under reduced pressure. 4 as a
white solid was obtained (59.6 g, 93.8% yield) with purity
1
>99.9% A (HPLC). H NMR (500 MHz, DMSO-d6): δ 1.40
(s, 9H), 1.91 (m, 1H), 2.01 (m, 1H), 2.75 (s, 3H), 3.02 (m,
1H), 3.11 (m, 1H), 3.28 (m, 2H), 4.70 (s, 1H), 7.44−7.29 (m,
3H), 7.92 (m, 2H), 10.27 (s, 2H). 13C NMR (126 MHz,
DMSO-d6): δ 26.96, 27.94, 28.08, 29.06, 43.41, 44.79, 54.03,
N
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
79.02, 114.46, 127.44, 128.95, 129.75, 137.50, 154.63, 170.09.
HRMS [M + H]+ m/z: 201.1603; theoretical, 201.1598.
2-Amino-4-{(3R)-3-[tert-butoxycarbonyl(methyl)amino]-
pyrrolidin-1-yl}-6-[(cyclopropylmethyl)amino]pyrimidin-1-
ium Benzoate (5). To a 1 L double-jacketed reactor equipped
with an impeller stirrer were charged 4 (194.7 g, 604.1 mmol,
1.2 equiv), dipotassium hydrogen phosphate (166.6 g, 956.4
mmol, 1.9 equiv), 3 (100.0 g, 503.4 mmol, 1.0 equiv), water
(200.0 g), and xylenes (400.0 g). The mixture was stirred
slowly and heated to gentle reflux (97−99 °C). The mixture
was aged for 20 h with fast stirring, then cooled to 60 °C in 2.5
h. A solution of potassium hydroxide (19.8 g, 352.4 mmol, 0.7
equiv) in water (100 g) was added to the mixture. The reaction
mixture was cooled to 40 °C in 1.5 h. The stirring was stopped,
and the lower aqueous phase was removed. Water (200 g) was
slowly added and the mixture was stirred at 40 °C for 30 min.
The mixture was filtered on a pressure filter and the solids were
rinsed twice with xylenes (100 g) and with water (100 g). The
solids were dried at 85 °C under reduced pressure. 5 was
isolated as a white solid (240.5 g, 98.6% yield) with purity
(m, 1H), 2.68 (s, 3H), 2.95 (d, J = 7.1 Hz, 2H), 3.46 (d, J =
8.1 Hz, 1H), 3.49−3.58 (m, 1H), 3.58−3.66 (m, 1H), 3.71−
3.80 (m, 1H), 3.84−3.95 (m, 1H), 4.87 (s, 1H). 13C NMR
(101 MHz, D2O): δ 2.97, 9.31, 27.31, 31.66, 44.89, 46.56,
48.70, 57.66, 71.79, 153.56, 154.85, 158.58. HRMS [M + H]+
m/z: 263.1978; theoretical, 263.1979.
2-Amino-6-[(cyclopropylmethyl)amino]-4-[(3R)-3-
(methylazaniumyl)pyrrolidin-1-yl]pyrimidin-1-ium (2R,3R)-
2,3-Dihydroxybutanedioate−Methanol−Water (1/x/y) (13).
To a 250 mL double-jacketed reactor equipped with an
impeller stirrer were charged 6 (45.8 g, 134.2 mmol, 1.0 equiv)
and butanol (179.0 g). A solution of sodium hydroxide (11.3 g,
282.9 mmol, 2.1 equiv) in water (125.0 g) was charged. The
solution was warmed up to 20 °C and the phases were split.
The upper organic layer was extracted twice with a 5% w/w
aqueous solution of sodium chloride at 20 °C. The mixture was
heated to 65 °C and distilled under reduced pressure to ca. 150
g. The residue was filtered to remove residual inorganics and
diluted with 1-butanol (115.0 g), and this solution was added
dropwise to a solution of L-(+)-tartaric acid (20.14 g, 134.2
mmol, 1.0 equiv) in methanol (342.4 g) and water (82.2 g) at
60 °C. After charging half of the solution, the addition was
stopped and the mixture was seeded and stirred at 60 °C for 1
h. The addition was resumed and completed in 1 h. The
resulting suspension was cooled to 20 °C in 4 h, stirred for 5 h,
and then cooled to 5 °C in 1 h. The suspension was stirred for
1 h and then filtered on a pressure filter. The solids were rinsed
twice with cold methanol (130 g) and dried at 55 °C under
reduced pressure. 13 was obtained as a white solid (53.8 g,
1
99.9% A (HPLC). H NMR (500 MHz, DMSO-d6): δ 0.21−
0.16 (m, 2H), 0.47−0.37 (m, 2H), 1.05−0.95 (m, 1H), 1.41
(s, 9H), 2.06−1.95 (m, 2H), 2.72 (s, 3H), 3.05 (m, 2H),
3.32−3.17 (m, 2H), 3.50 (m, 2H), 4.62 (s, 1H), 4.85 (s, 1H),
6.14 (s, 2H), 6.73 (s, 1H), 7.45 (m, 2H), 7.60−7.50 (m, 1H),
7.98−7.92 (m, 2H). 13C NMR (101 MHz, DMSO-d6): δ 3.66,
3.74, 11.30, 28.10, 28.32, 28.53, 28.71, 29.35, 45.19, 45.39,
48.20, 54.17, 72.91, 79.43, 128.58, 128.62, 128.81, 129.45,
129.68, 132.22, 133.86, 155.24, 160.32, 161.16, 161.52, 169.44.
HRMS [M + H]+ m/z: 363.2505; theoretical, 363.2503.
2-Amino-6-[(cyclopropylmethyl)amino]-4-[(3R)-3-
(methylazaniumyl)pyrrolidin-1-yl]pyrimidin-1-ium Dichlor-
ide (6). To a 1 L double-jacketed reactor equipped with an
impeller stirrer were charged 5 (120 g, 184.0 mmol, 1.0 equiv),
water (240 g), and toluene (480.0 g). The mixture was heated
to 45 °C in 20 min. 30% aqueous sodium hydroxide (65.0 g,
487.5 mmol, 2.65 equiv) was charged until pH > 12, and the
mixture was stirred until all solids dissolved. The aqueous
phase was removed and the organic phase was extracted twice
with water (120.0 g) at 45 °C. The solution was concentrated
under reduced pressure until the distillate flow ceased and the
residue was diluted with methanol (240 g). The distillation was
resumed until the distillate flow ceased. The residue was
diluted with methanol to a concentration of 33.5% w/w of 5
free base (198.4 g, solution A).
1
91.1% yield) with purity >99.95% A (HPLC). H NMR (500
MHz, DMSO-d6): δ 0.34−0.03 (m, 2H), 0.56−0.34 (m, 2H),
1.23−0.82 (m, 1H), 2.17−1.98 (m, 1H), 2.34−2.17 (m, 1H),
2.55 (s, 3H), 3.12−2.93 (m, 2H), 3.19 (s, 2H), 3.43−3.29 (m,
1H), 3.57−3.43 (m, 2H), 3.79−3.57 (m, 2H), 4.04 (s, 2H),
4.89 (s, 1H), 6.72 (br s, 10H). 13C NMR (101 MHz, DMSO-
d6): δ 3.79, 11.06, 27.85, 31.79, 44.88, 45.55, 49.07, 49.26,
57.18, 72.47, 73.23, 158.71, 159.75, 160.67, 176.68. HRMS [M
+ H]+ m/z: 263.1978; theoretical, 263.1979.
ASSOCIATED CONTENT
* Supporting Information
■
sı
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.1c00077.
Analytical methods and details about screenings of
reaction conditions (PDF)
To a separate 1 L double-jacketed reactor equipped with an
impeller stirrer was charged a 12.5% w/w methanolic solution
of HCl (160.5 g, 550.0 mmol, 3.0 equiv) freshly prepared by
dissolving gaseous HCl in methanol. The solution was heated
to 50 °C and solution A was charged in 6 h. Seeds were
charged and the mixture was stirred for 4 h at 50 °C. The
mixture was cooled to 35 °C in 1 h. The mixture was heated to
45 °C in 20 min, stirred for 5 min, and cooled to 35 °C in 100
min. The mixture was stirred at 35 °C for 1 h. The heating−
cooling cycle was repeated 3 times. Afterwards, the mixture
was cooled to −5 °C in 4 h, then TBME (136.0 g) was charged
in 30 min. The suspension was stirred for 1 h. The solids were
collected on a pressure filter and rinsed with a cold solution of
TBME (187.5 g) and methanol (62.5 g). The solids were dried
at 65 °C under reduced pressure. 6 was obtained as a white
solid (55.63 g, 93.4% yield) with purity >99.95% A (HPLC).
1H NMR (500 MHz, D2O): δ 0.04−0.21 (m, 2H), 0.38−0.51
(m, 2H), 0.89−1.01 (m, 1H), 2.08−2.26 (m, 1H), 2.32−2.48
AUTHOR INFORMATION
Corresponding Author
■
Ernesto Santandrea − Chemical and Analytical Development,
Novartis Pharma AG, 4002 Basel, Switzerland;
orcid.org/0000-0002-1363-3486;
Email: ernesto.santandrea@novartis.com
Authors
Christine Waldraff − Novartis Pharma AG, 4002 Basel,
Switzerland; orcid.org/0000-0001-5383-3231
Gilles Gerber − Novartis Pharma AG, 4002 Basel,
Switzerland
Maël Moreau − Novartis Pharma AG, 4002 Basel,
Switzerland
Pascal Beney − Novartis Pharma AG, 4002 Basel,
Switzerland
O
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
analogues of 1,4-benzodiazepine-based CCK 2 antagonists that
display high selectivity over CCK 1 receptors. J. Med. Chem. 2006,
49, 2253−2261. (b) Yoshimura, A.; Luedtke, M. W.; Zhdankin, V. V.
(Tosylimino)phenyl-λ3-iodane as a reagent for the synthesis of methyl
carbamates via Hofmann rearrangement of aromatic and aliphatic
carboxamides. J. Org. Chem. 2012, 77, 2087−2091. (c) Kumar, K.;
Awasthi, D.; Lee, S.-Y.; Cummings, J. E.; Knudson, S. E.; Slayden, R.
A.; Ojima, I. Benzimidazole-based antibacterial agents against
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.1c00077
Author Contributions
The manuscript was written through contributions of all the
authors. All the authors have given approval to the final version
of the manuscript.
Francisella tularensis. Bioorg. Med. Chem. 2013, 21, 3318−3326.
.
Notes
Some examples of reaction of isocyanates with t-butoxide: (d) Carter,
P. H.; Cherney, R. J.; Batt, D. G.; Brown, G. D.; Duncia, J. V.;
Gardner, D. S.; Yang, M. G. Substituted cycloalkylamine derivatives as
modulators of chemokine receptor activity. U.S. Patent
20,050,054,626 A1, 2005. (e) Fevig, J. M.; Feng, J.; Ahmad, S.
Pyrrolo(oxo)quinolines as 5HT ligands. U.S. Patent 20,060,014,778
A1, 2006.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors wish to thank Dr. Karine Lafleur for the support
and fruitful discussions upon scale-up. The authors also wish to
thank Dr. Serge Konrad and Luana Arrigo for the development
of the analytical methods. Finally, the authors wish to thank
Dr. Aldo Hörmann for the critical review of the manuscript.
(11) Prashad, M.; Har, D.; Hu, B.; Kim, H.-Y.; Girgis, M. J.;
̌
Chaudhary, A.; Repic, O.; Blacklock, T. J.; Marterer, W. Process
development of a large-scale synthesis of TKA731: a tachykinin
receptor antagonist. Org. Process Res. Dev. 2004, 8, 330−340.
(12) The predictions in the statistical system for impurities 36 and
47 showed unequivocal positive structural alert for mutagenicity, due
to the aromatic amine moiety. Following expert review, and
considering the loose similarity to the DS, the impurities were
subjected to Ames testing.
(13) (a) Hu, J.; Wang, X.; Chen, L.; Huang, M.; Tang, W.; Zuo, J.;
Liu, Y.-C.; Shi, Z.; Liu, R.; Shen, J.; Xiong, B. Design and discovery of
new pyrimidine coupled nitrogen aromatic rings as chelating groups
of JMJD3 inhibitors. Bioorg. Med. Chem. Lett. 2016, 26, 721−725.
(b) Saluja, A. K.; Tiwari, M.; Vullo, D.; Supuran, C. T. Substituted
Benzene Sulfonamides Incorporating 1,3,5-triazinyl Moieties Potently
Inhibit Human Carbonic Anhydrases II, IX and XII. Bioorg. Med.
Chem. Lett. 2014, 24, 1310−1314. (c) Du, W.; Baiazitov, R.; Lee, C.-
S.; Moon, Y.-C.; Paget, S. D.; Ren, H.; Sydorenko, N.; Wilde, R. G.
Substituted pyrimidine BMI-1 inhibitors. WO 2015030847 A1, 2015.
(d) Ioannidis, S.; Lamb, M. L.; Wang, T.; Almeida, L.; Block, M. H.;
Davies, A. M.; Peng, B.; Su, M.; Zhang, H.-J.; Hoffmann, E.; Rivard,
C.; Green, I.; Howard, T.; Pollard, H.; Read, J.; Alimzhanov, M.;
Bebernitz, G.; Bell, K.; Ye, M.; Huszar, D.; Zinda, M. Discovery of 5-
Chloro-N2-[(1S)-1-(5-fluoropyrimidin-2-yl)ethyl]-N4-(5-methyl-1H-
pyrazol-3-yl)pyrimidine-2,4-diamine (AZD1480) as a Novel Inhibitor
REFERENCES
■
(1) (a) Werfel, T.; Layton, G.; Yeadon, M.; Whitlock, L.; Osterloh,
I.; Jimenez, P.; Liu, W.; Lynch, V.; Asher, A.; Tsianakas, A.; Purkins,
L. Efficacy and safety of the histamine H4 receptor antagonist ZPL-
3893787 in patients with atopic dermatitis. J. Allergy Clin. Immunol.
2019, 143, 1830−1837. (b) Paller, A. S.; Kabashima, K.; Bieber, T.
Therapeutic pipelin for atopic dermatitis: End of the drought? J.
Allergy Clin. Immunol. 2017, 140, 633−643. (c) Werfel, T. Novel
systemic drugs in treatment of atopic dermatitis: results from phase II
and phase III studies published in 2017/2018. Curr. Opin. Allergy Clin.
Immunol. 2018, 18, 432−437. (d) Bell, A.; Lane, C. A. L.; Mowbray,
C. E.; Selby, M. D.; Swain, N. A.; Williams, D. H. Pyrimidine
derivatives. WO 2007072163 A2, 2007. (e) Zhu, Z.; Barker, H.;
Yeadon, M.; Liu, W. Pyrimidine compositions, ultra-pure composi-
tions and salts thereof, methods of making the same, and methods of
using the same for treating histamine H4 receptor (H4) mediated
diseases and conditions. WO 2017072131 A1, 2017.
(2) (a) REACH webpage for DMF. https://echa.europa.eu/
documents/10162/64875a92-e9c1-3d65-6722-0fab6c7283b3 (last ac-
cessed on March 18, 2021). (b) REACH webpage for DME. https://
echa.europa.eu/candidate-list-table/-/dislist/details/
0b0236e1807daf82 (last accessed on March 18, 2021).
(3) Chen, B.; Wang, L.; Gao, S. Recent Advances in Aerobic
Oxidation of Alcohols and Amines to Imines. ACS Catal. 2015, 5,
5851−5876.
(4) Liu, L.; Zhang, S.; Fu, X.; Yan, C.-H. Metal-free aerobic oxidative
coupling of amines to imines. Chem. Commun. 2011, 47, 10148−
10150.
(5) Tomilov, Y. V.; Kostistyn, A. B.; Shulishov, E. V.; Nefedov, O.
M. Palladium(II)-Catalyzed cyclopropanation of simple allyloxy and
allylamino compounds and of 1-oxy-1,3-butadienes with diazo-
methane. Synthesis 1990, 246−248.
(6) Teasdale, A.; Elder, D.; Chang, S.-J.; Wang, S.; Thompson, R.;
Benz, N.; Sanchez Flores, I. H. Rik assessment of genotoxic impurities
in new chemical entities: strategies to demonstrate control. Org.
Process Res. Dev. 2013, 17, 221−230.
(7) Hoffmann, G. R. Genetic effects of dimethyl sulfate, diethyl
sulfate and related compounds. Mutat. Res. 1980, 75, 63−129.
(8) Wan, Y.; Alterman, M.; Larhed, M.; Hallberg, A. Dimethylfor-
mamide as a Carbon Monoxide Source in Fast Palladium-Catalyzed
Aminocarbonylations of Aryl Bromides. J. Org. Chem. 2002, 67,
6232−6235.
(9) Tom, N. J.; Simon, W. M.; Frost, H. N.; Ewing, M. Deprotection
of a primary Boc group under basic conditions. Tetrahedron Lett.
2004, 45, 905−906.
of the Jak/Stat Pathway. J. Med. Chem. 2011, 54, 262−276.
.
Published after this work was carried out: (e) Chen, A.; Wu, D.-L.;
Shi, J.; Narva, S.; Zhao, X.-Y.; Wu, Y.-L.; Zhang, W. Design, Synthesis
and Biological Evaluation of 2-methyl-(1,1′-biphenyl)-pyrimidine
Conjugates. Bioorg. Med. Chem. Lett. 2020, 30, 127328−127351.
(14) (a) Meduna, S. P.; Savall, B. M.; Cai, H.; Edwards, J. P.;
Thurmond, R. L.; McGovern, P. M. Triamino pyrimidines and
pyridines as histamine H4 receptor modulators. Bioorg. Med. Chem.
Lett. 2011, 21, 3113−3116. (b) Shen, J.; Xiong, B.; Hu, J.; Geng, M.;
Li, J. Class I histone demethylase JMJD3 inhibitor and its preparation
method and application. CN 106608869 A, 2017.
(15) pKa measured in aqueous solutions: formic acid: 3.65, benzoic
́
́
acid: 4.20, acetic acid: 4.75. Carre, B.; Devynck, J. Proprietes acide-
́
base d’acide carboxyliques, d’amines et d’acides amines dans le
́
trifluoroethanol. Bull. Soc. Chim. Fr. 1981, 1, I309−I312.
(16) Calculated with the NRTL model in the utility “vapour−liquid
and liquid−liquid equilibrium prediction” (VLLE) from DynoChem
5, Scale-up Systems Ltd.; Dublin, Ireland.
(17) Modelled with DynoChem 5, Scale-up Systems Ltd.; Dublin,
Ireland.
(10) Some examples of reactions of t-butanol and other alcohols
with isocyanates: (a) McDonald, I. M.; Austin, C.; Buck, I. M.;
Dunstone, D. J.; Griffin, E.; Harper, E. A.; Hull, R. A. D.; Kalindjian,
S. B.; Linney, I. D.; Low, C. M. R.; Pether, M. J.; Spencer, J.; Wright,
P. T.; Adatia, T.; Bashall, A. Novel, achiral 1,3,4-benzotriazepine
P
https://doi.org/10.1021/acs.oprd.1c00077
Org. Process Res. Dev. XXXX, XXX, XXX−XXX