Development of a Scalable Negishi Cross-Coupling Process for the Preparation of 2-Chloro-5-(1-(tetrahydro-2 H-pyran-2-yl)-1 H-pyrazol-5-yl)aniline

Article abstract of DOI:10.1021/acs.oprd.0c00414

A scalable synthesis of 2-chloro-5-(1-(tetrahydro-2H-pyran-2-yl)-1H-pyrazol-5-yl)aniline (1), a key intermediate in the synthesis of an immuno-oncology asset, is described. A Negishi cross-coupling between in situ generated heteroaryl zinc reagent 4 and 5-bromo-2-chloroaniline (3) catalyzed by Pd(Xantphos)Cl2 enabled the construction of the key aryl-heteroaryl bond. A scalable first-generation process was developed that delivered 1 in multikilogram quantities. Building upon knowledge from the initial process, a more efficient workup and isolation procedure was developed that controlled levels of residual Pd and Zn to consistent levels that were acceptable for downstream processing. The high-yielding optimized process offers streamlined metals remediation, a 30% reduction in the number of unit operations, and a 34% reduction in process mass intensity (PMI) compared to the initial process.

Full text of DOI:10.1021/acs.oprd.0c00414

pubs.acs.org/OPRD
Article
Development of a Scalable Negishi Cross-Coupling Process for the
Preparation of 2‑Chloro-5-(1-(tetrahydro‑2H‑pyran-2-yl)‑1H‑pyrazol-
5-yl)aniline
Candice L. Joe,* Bahar Inankur, James Chadwick, Sha Lou, Jeffrey Nye, Neil A. Strotman,
and Albert J. DelMonte
Cite This: https://dx.doi.org/10.1021/acs.oprd.0c00414
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: A scalable synthesis of 2-chloro-5-(1-(tetrahydro-2H-pyran-2-yl)-1H-pyrazol-5-yl)aniline (1), a key intermediate in
the synthesis of an immuno-oncology asset, is described. A Negishi cross-coupling between in situ generated heteroaryl zinc reagent
4 and 5-bromo-2-chloroaniline (3) catalyzed by Pd(Xantphos)Cl2 enabled the construction of the key aryl−heteroaryl bond. A
scalable first-generation process was developed that delivered 1 in multikilogram quantities. Building upon knowledge from the
initial process, a more efficient workup and isolation procedure was developed that controlled levels of residual Pd and Zn to
consistent levels that were acceptable for downstream processing. The high-yielding optimized process offers streamlined metals
remediation, a 30% reduction in the number of unit operations, and a 34% reduction in process mass intensity (PMI) compared to
the initial process.
KEYWORDS: Zinc, Negishi, Pd-catalyzed, metals remediation, pyrazole, hexyllithium
reported by Sames and co-workers.2 Small scale experiments
INTRODUCTION
■
using a 3-fold excess of 2 revealed that the transformation was,
During the development of an immuno-oncology asset, 2-
indeed, selective for the desired regioisomer, albeit in an 85:15
chloro-5-(1-(tetrahydro-2H-pyran-2-yl)-1H-pyrazol-5-yl)-
ratio3 relative to other mono- and bis-functionalized pyrazole
aniline (1) was identified as a key intermediate in the first step
side products (Scheme 1). As we anticipated it could be
of a convergent synthetic route. A robust chemical process to
difficult to purge high levels of these structurally similar side
deliver 1 in multikilogram quantities was required to support
products (vide infra) in subsequent steps, we focused efforts on
downstream process development. While there are many
identifying milder conditions to selectively functionalize the 5-
possible approaches to prepare 1, we identified the aryl−
position of 2. In particular, pyrazole 2 is known to undergo
heteroaryl bond as the key disconnection (Figure 1). To
selective deprotonation at the 5-position at −78 °C with n-
achieve this goal, the appropriate protected pyrazole and 2,5-
butyllithium (n-BuLi), followed by trapping with an alkyl
bis-halogenated aniline starting materials were required. We
electrophile.4,5 We hypothesized that transmetalation of the
identified 1-(2-tetrahydropyranyl)-1H-pyrazole 2 as a com-
heteroaryl lithium of 2 to an appropriate zinc salt would
petent coupling partner. The tetrahydropyranyl (THP) group
generate a heteroaryl zinc reagent in situ that could participate
is compatible with downstream chemistry and can easily be
in Negishi cross-coupling with 3 to generate 1.6,7 Proof-of-
removed later in the synthesis. Although the potential
concept experiments suggested that this was the case, with 1
introduction of added analytical complexity from using the
formed selectively and no observable overfunctionalization of
chiral THP protecting group should always be considered, this
the aniline ring (Scheme 1).
was not a confounding factor in the preparation of 1. 5-Bromo-
2-chloroaniline 3 was chosen as the electrophile, as it allowed
for selective functionalization at the 5-position and provided a
handle for subsequent functionalization at the 2-position. Most
importantly, both 2 and 3 are readily available commercial
starting materials. Herein, we disclose a streamlined and high
yielding Negishi1 cross-coupling process for the preparation of
1.
HIGH-THROUGHPUT CATALYST SCREENING
■
After achieving proof-of-concept for the Negishi coupling, a
high-throughput catalyst screen was initiated (24 μmol scale).
Follow-up experiments revealed that deprotonation of 2 with
n-BuLi and transmetalation to ZnCl2 occurred at −10 °C,
Special Issue: Celebrating Women in Process Chem-
C−C BOND FORMING STRATEGIES
istry
■
Our development work focused on the construction of the key
C−C bond using metal-catalyzed cross coupling methods
(Scheme 1). Of particular interest was the regioselective
palladium-catalyzed direct C−H arylation of pyrazoles
Received: September 18, 2020
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 1. Retrosynthetic analysis of 1.
Scheme 1. Cross-coupling Approaches to 1
Table 1. Selected Ligand Screening Results from HTE Screen
a
b
Entry
Ligand
PPh3
1 (AP)
1
2
3
4
5
6
7
8
42
34
81
87
67
74
62
84
12
77
79
63
64
PCy3·HBF4
P(t-Bu)3·HBF4
X-Phos
S-Phos
RuPhos
Cy-JohnPhos
Xantphos
Cy-Xantphos
DPEPhos
DCEPhos
DPPF
9
10
11
12
13
DPPE
a
b
The ligand/metal (L/M) ratio for monodentate phosphine ligands was 2.2:1, and the L/M ratio for bidentate ligands was 1.1:1. Area percent
(AP) was determined by UPLC-MS.
which obviates the need to achieve cryogenic temperatures on
scale. Using this modified procedure, heteroaryl zinc reagent 4
was generated in a single batch that exists as a slurry upon
warming to 20 °C. The addition of 20 vol % N-Methyl-2-
pyrrolidone (NMP) was necessary to fully dissolve heteroaryl
zinc 4, thus facilitating uniform dosing into each screening
vial.8 Screening 36 ligands with 2 mol % [Pd(allyl)Cl]2
revealed a number of promising ligands with selected results
shown in Table 1.9 The best ligands included P(t-Bu)3·HBF4
(entry 3), X-Phos (entry 4), and Xantphos (entry 8), all of
B
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 2. Comparison at 1.5 mol % Catalyst Loading
Scheme 3. Initial Negishi Process To Generate 1
which provided Negishi product 1 in >80 area percent (AP)
with the remainder of the mass balance as residual starting
materials.
zincate formation and manifest in the Negishi screening results.
The use of 2-Me-THF during the formation of heteroaryl zinc
4 would obviate the need for a ternary reaction solvent
mixture. While 2-Me-THF and THF both provided a high AP
of 1 in a similarly fast time frame, the heteroaryl lithium
intermediate was a thick slurry in 2-Me-THF that made
agitation challenging. By contrast, the heteroaryl lithium slurry
is thinner and more easily suspended in THF, making it a
better choice of reaction solvent for scale up purposes with
overhead stirring.
A further evaluation of catalyst activity was undertaken on
650 μmol scale using P(t-Bu)3·HBF4/[Pd(allyl)Cl]2 (2.2:1 L/
M ratio), X-Phos/[Pd(allyl)Cl]2 (2.2:1 L/M ratio), and
Pd(Xantphos)Cl2 at 50 °C. All catalyst systems exhibited fast
kinetics in the Negishi coupling to form 1 at 50 °C, with all
reactions reaching completion (>95 AP 1) within 30 min
(Scheme 2). Despite exhibiting the slowest kinetics of the
three catalyst combinations, pre-complexed Pd(Xantphos)Cl2
was pursued to minimize the number of required unit
operations during scale-up. In particular, the use of this
catalyst avoids a metal/ligand pre-complexation step that
would require a separate catalyst preparation vessel followed by
an inert transfer to the in situ generated solution of 4.
The reaction is a two-step process: in situ formation of
heteroaryl zinc 4 followed by Negishi coupling with 3. In the
context of high-throughput screening, it was more practical to
generate 4 in a single solvent system, THF/hexane/2-Me-THF
(where hexane and 2-Me-THF were introduced by n-BuLi and
ZnCl2, respectively), to compare catalysts and ligands. This
avoids any inconsistencies that could originate during the
INITIAL PROCESS
■
With the palladium catalyst selected, we began to develop an
enabling process to deliver kilogram quantities of 1 to support
downstream process development (Scheme 3). Initial catalyst
screening was performed using 2 as the limiting reagent.
However, we quickly realized the benefits of making 3 the
limiting agent (Scheme 3) due to the more efficient purge of
residual pyrazole 2 compared to the rejection of 3 during the
crystallization (vide infra).10
Process safety and environmental concerns led us to replace
n-BuLi with n-hexyllithium (HexLi) for scale up purposes,11
with the deprotonation of pyrazole 2 also proceeding smoothly
C
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
at −10 °C with HexLi. The heteroaryl lithium intermediate
was found to be a slurry, and the deprotonation event
exothermic with an adiabatic temperature rise of 48 °C.12 A 30
min slow addition of HexLi to a −10 °C solution of pyrazole 2
in THF could control the exotherm. In doing so, the internal
temperature was never greater than 5 °C and in situ FTIR
spectroscopy confirmed that the consumption of pyrazole 2 is
essentially addition-controlled. The addition of the ZnCl2
solution to the heteroaryl lithium slurry was also exothermic,
exhibiting an adiabatic temperature rise of 35 °C. Slow
addition of the ZnCl2 solution over 30 min mitigated the safety
risk of this exotherm and generated heteroaryl zinc species 4
safely.
The Negishi coupling reached full conversion within 1 h
using 1.4 equiv of ZnCl2 (1.9 M solution in 2-Me-THF). Full
conversion to 1 was also observed within 1 h when
substoichiometric amounts of the ZnCl2 solution (0.7 equiv)
were used, albeit with slightly slower reaction kinetics (Figure
2).7i,13 Solid ZnCl2 and ZnBr2 were also suitable Zn precursors
Once the ZnCl2 solution charge was complete, the slurry of
in situ generated 4 was warmed to 20 °C. Aniline 3 and
Pd(Xantphos)Cl2 were charged to the reactor containing the
heteroaryl zinc slurry with extra care taken to exclude oxygen.
The Negishi coupling began upon heating the contents of the
reactor to 50 °C and aging for 4 h, which consistently delivered
1 in >98% in-process yield.
The main side product observed at the end of reaction is 6,
which arises from transmetalation of HexLi to ZnCl2 and
coupling of the corresponding hexylzinc species 5 with 3 under
the Negishi coupling conditions (Scheme 4). Analysis of the
end-of-reaction HPLC purity reveals that hexyl-coupled
impurity 6 is typically formed in <1 AP. While lowering the
equivalents of HexLi and ZnCl2 relative to 2 suppressed the
formation of 6, we found that a small overcharge of these
reagents in the heteroaryl zinc formation step increased
robustness by accounting for slight deviations in the potency
of the HexLi and ZnCl2 solutions.14 Moreover, 6 is efficiently
purged under our crystallization conditions. Notably, no side
products arising from overfunctionalization or dehalogenation
of the aniline ring were detected under the Negishi reaction
conditions. The THP protecting group also remained intact
throughout the process, which further highlights the mild
reaction and workup conditions (vide infra).
An appropriate metals remediation strategy is crucial for the
development of metal-catalyzed cross-coupling processes; in
the case of a Negishi coupling, metals removal must be
particularly robust, requiring the removal of both stoichio-
metric quantities of Zn and catalytic amounts of Pd.15 In
particular, high levels of Zn and Pd would challenge the
crystallization and potentially impact downstream trans-
formations and, ultimately, API quality. For the development
of our initial fit-for-purpose process, remediation of the two
metals was done in a stepwise fashion. Ideally, a Zn
remediation method would (1) quench the excess aryl zinc
reagent by protonolysis, (2) complex liberated zinc, and (3)
partition the Zn complex to the aqueous layer. The end-of-
reaction stream was washed with an aqueous solution of
ethylenediaminetetraacetic acid trisodium (EDTA·3Na) and
efficiently partitioned >98% of Zn to the aqueous layer, with
only 0.4 ppm remaining in the product-rich organic phase
(Table 2).7d Unfortunately, the EDTA·3Na wash was not an
Figure 2. Effect of ZnX2 source and stoichiometry on reaction
kinetics.
to access the heteroaryl zinc intermediate, with the Negishi
coupling reaching full conversion in a similar time frame to the
reaction using the ZnCl2 solution (Figure 2). The ZnCl2
solution in 2-Me-THF solution was preferred for scale up
purposes due to robustness concerns around suspending solid
zinc salts in the reactor early in the transmetalation process and
the potential challenges of reliably sourcing high quality
anhydrous ZnX2 in bulk quantities. The 2-Me-THF solution of
ZnCl2 has been consistently sourced with a KF < 1100 ppm,
thus mitigating the risk of heteroaryl lithium quenching due to
the presence of water. Since excess zinc was easily purged
during the workup (vide infra), the ZnCl2 charge was
maintained at 1.4 equiv to ensure process robustness.
a
Table 2. Metals Remediation in the Initial Process
End-of-reaction
stream
NAc Wash,
pH = 7
Analyte
EDTA·3Na Wash
Zn (ppm)
Pd (ppm)
27,886
325
0.4
338
0.4
103
a
Levels of residual metals in the product-rich organic phase were
determined by X-ray Fluorescence Spectroscopy (XRF).
efficient Pd scavenger, with no reduction in Pd levels observed
in the organic phase. A second wash with pH = 7 aqueous N-
Scheme 4. Formation of Hexyl-Coupled Impurity 6
D
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
acetylcysteine (NAc) reduced the Pd levels in the organic layer
by approximately 70%. An additional 13% brine wash removed
residual inorganics from the organic phase prior to isolation of
1.
Cross-coupled product 1 exhibited high solubility (>50 mg/
mL) in the majority of common solvents (Table 3). For
OPTIMIZED PROCESS
■
The successful scale up of the initial process allowed us to
identify various areas for improvement. For example, the
workup required high kg/kg loadings of EDTA·3Na due to the
limited solubility of EDTA·3Na in water. This led to large
V
min/Vmax swings of the process stream during extraction.
Moreover, the aqueous NAc wash proved to be inconsistent on
scale with respect to Pd remediation. The requirement for
three aqueous washes for Pd and Zn remediation also resulted
in longer-than-ideal processing times and higher PMI due to
the waste stream generation. We sought to develop an
improved process that addressed these drawbacks with a
particular focus on streamlining the workup (Scheme 5).
A more efficient metals remediation strategy would remove
Pd and Zn in a single workup step. To evaluate this possibility,
a survey of potential scavengers was carried out on the end-of-
reaction stream (Table 4). Quenching the end-of-reaction
Table 3. Selected Solubility Data for 1 and 3 at 20 °C
Boiling
point
Product 1
(mg/mL)
Aniline 3
(mg/mL)
Solvent
THF
66 °C
80 °C
83 °C
106 °C
111 °C
100 °C
98 °C
410
154
36
128
95
1301
1145
605
659
896
0.03
58
2-Me-THF
i-PrOH
CPME
Toluene
Water
0.003
1
Heptane
33% Toluene/
66% Heptane
33% CPME/
66% Heptane
100 °C
6
308
Table 4. Scavenger Screen for Pd and Zn Remediation18
98 °C
9
3
462
10
Scavenger
No wash
10% w/v NH4OAc(aq)
EDTA·3Na(aq)
Ethylene diamine(aq)
NAc(aq), pH = 8
SiliaMetS Thiol
Zn, ppm
27,886
4653
0.3
1.4
Pd, ppm
Notes
33% i-PrOH/66% Water
82 °C
325
401
392
218
104
348
a
Solvent ratios are reported in v/v.
Clean phase split
Clean phase split
Clean phase split
Emulsion
isolation, a swap to a solvent with a higher boiling point than
the ternary reaction solvent mixture (THF/2-Me-THF/
hexane) was desired to derisk variations in the ternary solvent
ratios from batch to batch depending on solvent losses during
the reaction and workup. A solvent swap to a higher boiling
solvent at a fixed volume post-workup would provide a
platform for a controlled crystallization without variability
arising from residual reaction solvents. We identified three high
boiling solvents where 1 exhibits lower solubility than the
reaction solvents: isopropanol (i-PrOH), cyclopentyl methyl
ether (CPME), and toluene. Water and heptane were the only
suitable anti-solvents identified based on low solubility of 1
(<5 mg/mL). Based on this solubility data, a number of
solvent/anti-solvent combinations16 were screened to identify
a system capable of isolating 1 as a crystalline solid while
purging 2 and any residual 3. A key criteria for solvent/anti-
solvent selection was a >30 mg/mL solubility difference
between 1 and 3. We ultimately pursued a toluene/heptane
crystallization for our enabling process due to its better purge
of the aniline 3 in the event of incomplete conversion.
Not detected
12667
Requires filtration
a
Levels of residual metals in the product-rich organic phase were
determined by XRF.
stream with 10% aqueous NH4OAc resulted in a clean phase
split but incomplete removal of Zn and Pd. While the efficient
removal of Pd and Zn was observed with a single pH 8 N-
acetylcysteine wash, this was accompanied by an emulsion that
made phase separation difficult. An aqueous ethylene diamine
wash efficiently lowered the Zn levels in the organic phase to
levels similar to that of EDTA·3Na.7c,17 As an added benefit,
3.2 kg/kg EDTA·3Na in 12 volumes of water was required for
efficient Zn remediation, whereas 1.6 kg/kg ethylene diamine
in 8 volumes of water was sufficient to lower Pd and Zn levels.
For this reason, the aqueous ethylene diamine wash was more
attractive from an efficiency and PMI perspective.
Quenching the reaction with an aqueous ethylene diamine
solution resulted in an adiabatic temperature rise of 15 °C.
While the exotherm was easily controlled by the slow addition
of the ethylene diamine quench solution over 90 min, some
solids formed in early stages of the addition. To avoid potential
hang-up of solids in the reactor, a reverse quench, whereby the
end-of-reaction stream was charged to the aqueous ethylene
diamine solution, was implemented for subsequent campaigns.
In the lab, 1 was consistently isolated in up to 80% yield
(>99.5 wt % and >99.4 LCAP purity) with acceptable levels of
residual metals (773 ppm Pd, 0.9 ppm Zn) on 10−50 g scale.
The scalability of this process was then demonstrated by the
successful delivery of three 2-kg batches of 1 with similar yields
and quality as observed on laboratory scale.
Scheme 5. Optimized Negishi Process To Generate 1
E
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
With a streamlined workup procedure in hand, the effect on
the crystallization and residual metals levels in the isolated
solids was further investigated (Table 5). The toluene/heptane
Table 6. Comparison of the Initial and Optimized Negishi
Processes
Optimized
Initial Process
3 extractions
Process
Improvement
Table 5. Lab Scale Comparison of Crystallization
Conditions on Quality of Isolated 1
Workup
1 extraction
Streamlined metals
remediation
Unit
Operations
7
5
30% reduction
Solvent System
Yield
Purity (LCAP)
Pd, ppm
Zn, ppm
Toluene/heptane
i-PrOH/water
82%
88%
99.4
99.6
1665
171
205
5
Purity
99.4
99.4%
79
99.6
99.7%
52
Good quality control
Good quality control
34% reduction
(LCAP)
Potency
(wt%)
PMI
crystallization gave 1 in comparable yield and purity to the
initial process; however, residual Pd and Zn levels in isolated
material were significantly higher. Based on the solubility data
for 1 (Table 3), we expected that an i-PrOH/water
crystallization would provide 1 with low product losses to
the mother liquor and is differentiated in its solvent polarity
from the toluene/heptane system. Performing an i-PrOH/
water crystallization after the aqueous ethylene diamine wash
provided 1 in high yield and quality and, most importantly,
with low levels of residual metals. Additionally, the toluene/
heptane and i-PrOH/water crystallization conditions provided
the same crystal form of 1. Analysis of the mother liquor and
subsequent washes revealed that i-PrOH/water was capable of
further purging trace metals post-workup whereas toluene/
heptane was not. Drying the wet cake overnight under vacuum
at 50 °C ensured that residual solvents and ethylene diamine
were removed. The control of residual ethylene diamine is an
important quality attribute in the dry cake of 1, as it can act as
a catalyst poison in subsequent metal-catalyzed steps. The
isolation procedure provided consistent control over the
ethylene diamine levels to <0.05 wt % in the dry cake.
THF) were purchased from Sigma-Aldrich. UPLCMS analysis
was performed using a Waters Acquity BEH Shield RP-18
column (1.7 μm, 2.1 mm × 50 mm) with detection by UV at
220 nm and low resolution mass spectrometry detection
(positive ion mode) with a Shimadzu LCMS-2020 mass
spectrometer. HPLC analysis was performed using a Supelco
Ascentis Express C-18 column (2.7 μm, 4.6 mm × 50 mm)
with UV detection at 220 nm. High-resolution mass
spectrometry (HRMS) was performed on an Agilent 6230B
TOF mass spectrometer. Data are presented as follows:
chemical shift (ppm), multiplicity (s = singlet, d = doublet, t =
triplet, m = multiplet, br = broad), coupling constant J (Hz),
and integration. Trace metals analysis was carried out using
two techniques: (1) X-ray fluorescence spectroscopy (XRF) on
a Malvern Panalytical Epsilon 1 spectrometer and (2)
Inductively coupled plasma atomic emission spectroscopy
(ICP-OES) using a ThermoFisher iCap 7400 spectrometer.
2-Chloro-5-(1-(tetrahydro-2H-pyran-2-yl)-1H-pyrazol-
5-yl)aniline (1). Initial Process. A solution of 1-(tetrahydro-
2H-pyran-2-yl)-1H-pyrazole 2 (479 g, 3.1 mol, 1.3 equiv) in
THF (4 L, 8 L/kg) was inerted via subsurface N2 sparging and
cooled to −10 °C. A solution of hexyllithium in hexane (1.4 L,
2.3 M, 3.2 mol, 1.35 equiv) was charged over 30 min such that
the internal temperature remained below 5 °C. After aging the
slurry for 20 min, a solution of ZnCl2 in 2-methyltetra-
hydrofuran (1.7 L, 1.9 M, 3.4 mol, 1.4 equiv) was added
portionwise such that the internal temperature remained below
5 °C. The slurry was warmed to 20 °C and inerted via
subsurface N2 sparging. 5-Bromo-2-chloroaniline 3 (500 g, 2.4
mol, 1.0 equiv), dichloro[9,9-dimethyl-4,5-bis(diphenyl-
phosphino)xanthene]palladium(II) (18.5 g, 24 mmol, 0.01
equiv), and THF (800 mL, 1.6 L/kg) were charged to the
reactor. The reaction mixture was heated to 50 °C for 4 h and
cooled to 20 °C once the reaction was judged to be complete
by HPLC analysis.
A solution of ethylenediaminetetraacetic acid trisodium (1.6
kg, 4.3 mol, 1.8 equiv) and water (6 L, 12 L/kg) was added to
the vessel. The mixture was stirred for 4 h, and the aqueous
layer was removed. The organic layer was washed with a
solution of N-acetyl-L-cysteine (625 g, 3.8 mol, 1.6 equiv),
potassium phosphate tribasic (675 g, 3.1 mol, 1.4 equiv), and
water (3 L, 6 L/kg) and aged at 20 °C for 14 h. The aqueous
layer was removed, and the organic layer was washed with 13%
brine (2.5 L, 5 mL/g). After the aqueous layer was discarded,
the organic layer was concentrated to 2.5 L (5 L/kg) followed
by constant-volume distillation and solvent exchange to
toluene. The temperature of the batch was adjusted to 45
°C followed by the addition of heptane (1.5 L, 3 L/kg) and
aging at 45 °C for 1 h to facilitate formation of a seed bed.
The optimized process was executed on 300-g scale,
delivering 1 in 88% yield and 99.6 LCAP purity. Importantly,
the combination of the ethylene diamine workup and i-PrOH/
water crystallization efficiently purged metals to an acceptable
level with only 426 ppm Pd and 0.8 ppm Zn remaining in the
isolated solids. These levels of Pd and Zn had no impact on the
subsequent steps or the purity of our drug candidate.
CONCLUSIONS
■
A scalable and robust Negishi coupling process was developed
that delivers 1 in consistent yield and quality that was
acceptable for downstream processing. This was enabled by a
variety of workflows: high-throughput experimentation to
identify the appropriate catalyst, solubility screening to find
an optimal crystallization solvent combination, and metal
scavenger screening to streamline the workup procedure.
Compared to the enabling process, the optimized process
offers a streamlined metals remediation, a 30% reduction in the
number of unit operations, and a 34% reduction in process
mass intensity19 (PMI, Table 6). This was accomplished by
using an aqueous ethylene diamine wash to remove residual Zn
and Pd and implementing an i-PrOH/water crystallization that
is capable of further purging trace metal impurities post-
workup.
EXPERIMENTAL SECTION
■
All operations were performed under a nitrogen atmosphere.
Starting materials, reagents, and solvents were used as-received
from commercial vendors. Standard benchtop techniques were
employed for handling air- and moisture-sensitive reagents.
Hexyllithium (2.3 M in hexane) and ZnCl2 (1.9 M in 2-Me-
F
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Heptane (3.5 L, 7 L/kg) was charged at 45 °C, and the batch
was aged at this temperature after the addition was complete.
The batch was cooled to 20 °C over 1 h and aged for an
additional 8 h. Filtration of the solids followed by cake washing
with 25% toluene in heptane (2.5 L, 5 L/kg) and heptane (2.5
mL, 5 L/kg) and drying under vacuum at 50 °C provided 1 as
a beige crystalline solid, 485 g (99.4 LCAP purity, 73% yield).
Optimized Process. A solution of 1-(tetrahydro-2H-pyran-
2-yl)-1H-pyrazole 2 (290 g, 1.9 mol, 1.3 equiv) in THF (2.4 L,
8 L/kg) was inerted via subsurface N2 sparging and cooled to
−10 °C. A solution of hexyllithium in hexane (620 g, 2.2 M,
2.7 mol, 1.4 equiv) was charged over 30 min such that the
internal temperature remained below 5 °C. After aging the
slurry for 20 min, a solution of ZnCl2 in 2-methyltetra-
hydrofuran (1.0 kg, 1.9 M, 2.7 mol, 1.4 equiv) was added
portionwise such that the internal temperature remained below
5 °C. The slurry was warmed to 20 °C and inerted via
subsurface N2 sparging. 5-Bromo-2-chloroaniline 3 (300 g, 1.5
mol, 1.0 equiv), dichloro[9,9-dimethyl-4,5-bis(diphenyl-
phosphino)xanthene]palladium(II) (11 g, 15 mmol, 0.01
equiv), and THF (0.60 L, 1.6 L/kg) were charged to the
reactor. The reaction mixture was heated to 50 °C for 4 h and
cooled to 20 °C once the reaction was judged to be complete
by HPLC analysis.
States; orcid.org/0000-0002-7167-2416;
Email: Candice.joe@bms.com
Authors
Bahar Inankur − Chemical Process Development, Bristol
Myers Squibb, New Brunswick, New Jersey 08903, United
States
James Chadwick − Chemical Process Development, Bristol
Myers Squibb, Moreton, Wirral CH46 1QW, United
Kingdom; orcid.org/0000-0003-4456-0117
Sha Lou − Chemical Process Development, Bristol Myers
Squibb, New Brunswick, New Jersey 08903, United States;
orcid.org/0000-0002-4867-4106
Jeffrey Nye − Chemical Process Development, Bristol Myers
Squibb, New Brunswick, New Jersey 08903, United States
Neil A. Strotman − Chemical Process Development, Bristol
Myers Squibb, New Brunswick, New Jersey 08903, United
States; orcid.org/0000-0002-5350-8735
Albert J. DelMonte − Chemical Process Development, Bristol
Myers Squibb, New Brunswick, New Jersey 08903, United
States; orcid.org/0000-0003-0645-3169
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00414
A solution of ethylene diamine (520 g, 8.7 mol, 6 equiv) in
water (2.4 L, 8 L/kg) was prepared in a separate reactor and
cooled to 20 °C. The reaction stream was charged to the
quench reactor over 1 h. After the mixture was stirred at 20 °C
for 4 h, the lower aqueous layer was removed. The organic
layer was concentrated to 1.5 L (5 L/kg) followed by constant-
volume distillation and solvent exchange to i-PrOH. The
temperature was adjusted to 50 °C followed by the addition of
water (1.5 L, 5 L/kg) and aging at 50 °C for 2 h to ensure seed
bed formation. The final portion of water (0.90 L, 3 L/kg) was
charged at 50 °C. The batch was cooled to 20 °C over 2 h and
aged at this temperature for an additional 8 h. Filtration of
solids followed by cake washing with 40% isopropanol in water
(1.5 L, 5 L/kg) and drying under vacuum at 50 °C provided
the title compound as a beige crystalline solid, 356 g (99.6
LCAP purity, 88% yield).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors would like to thank Christopher Jamison for initial
proof-of-concept studies for the Negishi cross-coupling, as well
as Merrill Davies for early analytical support. We thank Sharla
Wood for atomic spectroscopy support and Shasha Zhang for
process safety evaluations. Sergei Kolotuchin, Eric Simmons,
Matthew Winston, Rebecca Green, and Ling Zhang are
acknowledged for helpful discussions.
REFERENCES
■
(1) Negishi, E.; King, A. O.; Okukado, N. Selective Carbon-Carbon
Bond Formation Via Transition-Metal Catalysis 0.3. Highly Selective
Synthesis of Unsymmetrical Biaryls and Diarylmethanes by Nickel-
Catalyzed or Palladium-Catalyzed Reaction of Aryl Derivatives and
Benzylzinc Derivatives with Aryl Halides. J. Org. Chem. 1977, 42,
1821−1823.
1H NMR (500 MHz, chloroform-d) δ 7.58 (s, 1H), 7.32 (d,
J = 8.2 Hz, 1H), 6.92 (s, 1H), 6.83 (dd, J = 8.2, 1.2 Hz, 1H),
6.29 (s, 1H), 5.21 (dd, J = 10.2, 2.0 Hz, 1H), 4.21−4.10 (m,
3H), 3.62−3.57 (m, 1H), 2.61−2.53 (m, 1H), 2.09−2.04 (m,
1H), 1.84 (br d, J = 12.5 Hz, 1H), 1.78−1.72 (m, 1H), 1.62−
1.53 (m, 2H). 13C NMR (126 MHz, chloroform-d) δ 143.4,
143.0, 139.4, 130.0, 129.5, 119.6, 119.5, 116.1, 106.4, 84.1,
67.7, 29.7, 24.8, 22.9. HRMS (ESI+) Calculated (M + H)
278.1055, found 278.1064.
(2) Goikhman, R.; Jacques, T. L.; Sames, D. C-H Bonds as
Ubiquitous Functionality: A General Approach to Complex Arylated
Pyrazoles via Sequential Regioselective C-Arylation and N-Alkylation
Enabled by SEM-Group Transposition. J. Am. Chem. Soc. 2009, 131,
3042−3048.
(3) Ratio of 1- to 4-arylation and 4,5-bisarylation side products was
determined based on area percent using UPLCMS analysis of the end-
of-reaction stream at 220 nm. The ratios are not corrected for any
relative response differences between the compounds.
(4) Ahmed, B. M.; Mezei, G. Green protection of pyrazole, thermal
isomerization and deprotection of tetrahydropyranylpyrazoles, and
high-yield, one-pot synthesis of 3(5)-alkylpyrazoles. RSC Adv. 2015,
5, 24081−24093.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.0c00414.
(5) (a) Vedso, P.; Begtrup, M. Synthesis of 5-Substituted 1-
Hydroxypyrazoles through Directed Lithiation of 1-(Benzyloxy)-
Pyrazole. J. Org. Chem. 1995, 60, 4995−4998. (b) Gerard, A. L.;
Bouillon, A.; Mahatsekake, C.; Collot, V.; Rault, S. Efficient and
simple synthesis of 3-aryl-1H-pyrazoles. Tetrahedron Lett. 2006, 47,
4665−4669. (c) Gerard, A. L.; Mahatsekake, C.; Collot, V.; Rault, S.
A facile synthetic route to new pyrazoloisoindolones. Tetrahedron Lett.
2007, 48, 4123−4126. (d) Despotopoulou, C.; Klier, L.; Knochel, P.
Synthesis of Fully Substituted Pyrazoles via Regio- and Chemo-
Additional HTE screening data, NMR characterization
of 1, reaction calorimetry data (PDF)
AUTHOR INFORMATION
■
Corresponding Author
Candice L. Joe − Chemical Process Development, Bristol
Myers Squibb, New Brunswick, New Jersey 08903, United
G
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
selective Metalations. Org. Lett. 2009, 11, 3326−3329. (e) Roy, S.;
Roy, S.; Gribble, G. W., Metalation of Pyrazoles and Indazoles. In
Topics in Heterocyclic Chemistry; Gribble, G. W., Ed.; Springer-Verlag:
Berlin Heidelberg, 2012; Vol. 29, pp 155−260. (f) Kumpulainen, E.
T. T.; Pohjakallio, A. Selective Palladium-Catalyzed Direct C-H
Arylation of Unsubstituted N-Protected Pyrazoles. Adv. Synth. Catal.
2014, 356, 1555−1561. (g) Mistico, L.; Querolle, O.; Meerpoel, L.;
Angibaud, P.; Durandetti, M.; Maddaluno, J. Access to Silylated
Pyrazole Derivatives by Palladium-Catalyzed C-H Activation of a
TMS group. Chem. - Eur. J. 2016, 22, 9687−9692.
(6) Kristensen, J.; Begtrup, M.; Vedso, P. Preparation of 5-Acyl- and
5-Aryl-substituted 1-(benzyloxy)pyrazoles via directed Ortho-lithia-
tion/transmetalation and palladium catalyzed cross-coupling. Synthesis
1998, 1998, 1604−1608.
(14) An alternative strategy of increasing the equivalents of 2 relative
to the HexLi and ZnCl2 charges has also been successful at
suppressing the formation of hexyl-coupled impurity 6.
(15) (a) Konigsberger, K.; Chen, G. P.; Wu, R. R.; Girgis, M. J.;
Prasad, K.; Repic, O.; Blacklock, T. J. A practical synthesis of 6-[2-
(2,5-dimethoxyphenyl)ethyl]-4-ethylquinazoline and the art of
removing palladium from the products of Pd-catalyzed reactions.
Org. Process Res. Dev. 2003, 7, 733−742. (b) Garrett, C. E.; Prasad, K.
The art of meeting palladium specifications in active pharmaceutical
ingredients produced by Pd-catalyzed reactions. Adv. Synth. Catal.
2004, 346, 889−900. (c) Welch, C. J.; Albaneze-Walker, J.; Leonard,
W. R.; Biba, M.; DaSilva, J.; Henderson, D.; Laing, B.; Mathre, D. J.;
Spencer, S.; Bu, X. D.; Wang, T. B. Adsorbent screening for metal
impurity removal in pharmaceutical process research. Org. Process Res.
Dev. 2005, 9, 198−205. (d) Flahive, E. J.; Ewanicki, B. L.; Sach, N.
W.; O’Neill-Slawecki, S. A.; Stankovic, N. S.; Yu, S.; Guinness, S. M.;
Dunn, J. Development of an effective palladium removal process for
VEGF oncology candidate AG13736 and a simple, efficient screening
technique for scavenger reagent identification. Org. Process Res. Dev.
2008, 12, 637−645.
(7) (a) Manley, P. W.; Acemoglu, M.; Marterer, W.; Pachinger, W.
Large-scale Negishi coupling as applied to the synthesis of PDE472,
an inhibitor of phosphodiesterase type 4D. Org. Process Res. Dev.
2003, 7, 436−445. (b) Ragan, J. A.; Raggon, J. W.; Hill, P. D.; Jones,
B. P.; McDermott, R. E.; Munchhof, M. J.; Marx, M. A.; Casavant, J.
M.; Cooper, B. A.; Doty, J. L.; Lu, Y. Cross-coupling methods for the
large-scale preparation of an imidazole-thienopyridine: Synthesis of
[2-(3-methyl-3H-imidazol-4-yl)thieno[3,2-b]pyridin-7-yl]-(2-methyl-
1H-indol-5-yl)-amine. Org. Process Res. Dev. 2003, 7, 676−683.
(c) Denni-Dischert, D.; Marterer, W.; Banziger, M.; Yusuff, N.; Batt,
D.; Ramsey, T.; Geng, P.; Michael, W.; Wang, R. M. B.; Taplin, F.;
Versace, R.; Cesarz, D.; Perez, L. B. The synthesis of a novel inhibitor
of B-Raf kinase. Org. Process Res. Dev. 2006, 10, 70−77. (d) Perez-
Balado, C.; Willemsens, A.; Ormerod, D.; Aelterman, W.; Mertens, N.
Development of a concise scaleable synthesis of 2-chloro-5-(pyridin-
2-yl) pyrimidine via a Negishi cross-coupling. Org. Process Res. Dev.
2007, 11, 237−240. (e) Lu, B.; Li, G.; Roschangar, F.; Hossain, A.;
Herter, R.; Farina, V.; Senanayake, C. H. Development of a Practical
Negishi Coupling Process for the Manufacturing of BILB 1941, an
HCV Polymerase Inhibitor. In Transition Metal-Catalyzed Couplings in
Process Chemistry; Dunetz, J. M. J. R., Ed.; Wiley-VCH: Weinheim,
Germany, 2013; pp 105−120. (f) Acemoglu, M.; Baenziger, M.;
Marterer, W. Experiences with Negishi Couplings on Technical Scale
in Early Development. In Transition Metal-Catalyzed Couplings in
Process Chemistry; Dunetz, J., Dunetz, J. R., Eds.; Wiley-VCH:
Weinheim, Germany, 2013; pp 15−23. (g) Shu, L. H.; Wang, P.; Gu,
C.; Liu, W.; Alabanza, L. M.; Zhang, Y. C. An Efficient Large-Scale
Synthesis of a Naphthylacetic Acid CRTH2 Receptor Antagonist. Org.
Process Res. Dev. 2013, 17, 651−657. (h) Gontcharov, A.; Dunetz, J.
R. Rapid Enabling of Negishi Couplings for a Pair of mGluR5
Negative Allosteric Modulators. Org. Process Res. Dev. 2014, 18,
1145−1152. (i) Remarchuk, T.; Angelaud, R.; Askin, D.; Kumar, A.;
Thompson, A. S.; Cheng, H.; Reichwein, J. F.; Chen, Y. P.; St-Jean, F.
Manufacture of the PI3K beta-Sparing Inhibitor Taselisib. Part 1:
Early-Stage Development Routes to the Bromobenzoxazepine Core.
Org. Process Res. Dev. 2019, 23, 775−782.
(16) The boiling points of the solvent/antisolvent mixtures were
determined by Dynochem modelling using UNIFAC vapor−liquid
equilibra correlations.
(17) Wisniewski, S. R.; Stevens, J. M.; Yu, M.; Fraunhoffer, K. J.;
Romero, E. O.; Savage, S. A. Utilizing Native Directing Groups:
Synthesis of a Selective IKur Inhibitor, BMS-919373, via a
Regioselective C-H Arylation. J. Org. Chem. 2019, 84, 4704−4714.
(18) It is possible for metal levels to increase in the organic phase
post-workup due to a loss of organic mass to the aqueous wash.
(19) Jimenez-Gonzalez, C.; Ponder, C. S.; Broxterman, Q. B.;
Manley, J. B. Using the Right Green Yardstick: Why Process Mass
intensity Is Used in the Pharmaceutical Industry To Drive More
Sustainable Processes. Org. Process Res. Dev. 2011, 15, 912−917.
(8) The impact of NMP on the reactivity in the Negishi coupling
was not examined.
(9) See Table S1 in the Supporting Information for additional ligand
screening data with [Pd(allyl)Cl]2 and Pd(OAc)2.
(10) Pyrazole 2 is an oil, whereas aniline 3 is a solid. We had
confidience that, under controlled crystallization conditions, we would
be able to purge excess 2.
(11) (a) Rathman, T.; Schwindeman, J. A. Preparation, Properties,
and Safe Handling of Commercial Organolithiums: Alkyllithiums,
Lithium sec-Organoamides, and Lithium Alkoxides. Org. Process Res.
Dev. 2014, 18, 1192−1210. (b) Anderson, N. G. Practical Process
Research and Development: A Guide for Organic Chemists; Elsevier
Science: 2012.
(12) See Supporting Information for additional details.
(13) The fact that the reaction reaches full conversion with 1.4 and
0.7 equiv of ZnCl2 suggests that the both ArZnCl and Ar2Zn species
exist in solution and are capable of undergoing transmetalation to Pd.
We have been unable to distinguish between the two species using in
situ spectroscopic methods.
H
https://dx.doi.org/10.1021/acs.oprd.0c00414
Org. Process Res. Dev. XXXX, XXX, XXX−XXX