Exploratory process development and kilogram-scale synthesis of a novel oxazolidinone antibacterial candidate

Article abstract of DOI:10.1021/op500030v

A concise, environmentally benign, and cost-effective route was developed for the large-scale preparation of 1, a novel oxazolidinone antibacterial candidate. The key intermediate 2-(1-(2-fluoro-4-nitrophenyl)-1H-pyrazol-4-yl) pyridine 7 was prepared with high purity by mild deamination of the regioisomeric mixture 21. The mixture was prepared from a nucleophilic SNAr reaction by selective C-N coupling of the secondary amine functionality of 4-(pyridin-2-yl)-1H-pyrazol-3-amine 14 with 1,2-difluoro-4-nitrobenzene 10 in optimized conditions with the primary amine group remaining intact. The gaseous nitrogen release rate and reaction mixture temperature of the deamination step can be well controlled by altering the feeding manner, thereby providing safety guarantees. The optimized synthetic strategy of 1 with an overall yield of 27.6%, including seven sequential transformations by only five solid-liquid isolations, significantly improved the product separation workup. The strategy bypassed time-consuming and laborious procedures for any intermediate involved as well as for the final API. This study presents a process enabling the rapid delivery of a multikilogram quantity of API with high purity.

Full text of DOI:10.1021/op500030v

Article
pubs.acs.org/OPRD
Exploratory Process Development and Kilogram-Scale Synthesis of a
Novel Oxazolidinone Antibacterial Candidate
Tao Yang,†,§ Jia-Xiang Chen,†,§ Yiwei Fu,† Kaixiao Chen,‡ Jinyun He,‡ Weiwei Ye,† Zitai Sang,†
and Youfu Luo*,†
†State Key Laboratory of Biotherapy, West China Hospital, West China Medical School, Sichuan University, Chengdu, Sichuan,
610041, P.R. China
‡Sichuan Gooddoctor Pharmaceutical Co., Ltd., Chengdu, Sichuan, 610041, P.R. China
S
* Supporting Information
ABSTRACT: A concise, environmentally benign, and cost-effective route was developed for the large-scale preparation of 1, a
novel oxazolidinone antibacterial candidate. The key intermediate 2-(1-(2-fluoro-4-nitrophenyl)-1H-pyrazol-4-yl)pyridine 7 was
prepared with high purity by mild deamination of the regioisomeric mixture 21. The mixture was prepared from a nucleophilic
SNAr reaction by selective C−N coupling of the secondary amine functionality of 4-(pyridin-2-yl)-1H-pyrazol-3-amine 14 with
1,2-difluoro-4-nitrobenzene 10 in optimized conditions with the primary amine group remaining intact. The gaseous nitrogen
release rate and reaction mixture temperature of the deamination step can be well controlled by altering the feeding manner,
thereby providing safety guarantees. The optimized synthetic strategy of 1 with an overall yield of 27.6%, including seven
sequential transformations by only five solid−liquid isolations, significantly improved the product separation workup. The
strategy bypassed time-consuming and laborious procedures for any intermediate involved as well as for the final API. This study
presents a process enabling the rapid delivery of a multikilogram quantity of API with high purity.
compound 1 was necessary to expedite scale-up development
and quick delivery for regulatory toxicity and preclinical studies.
Retrosynthetic analysis of compound 1 was carried out
according to the general guidelines recommended by Warren3
(Figure 2).
INTRODUCTION
Linezolid (Figure 1) is the first representative of a new class of
oxazolidinone antibacterials. It has successfully been used to
■
The most efficient disconnection of A occurrs in the most
central area of the relevant starting molecule. This phenom-
enon leads to a highly convergent route to compound 1, which
meets our requirement of “scalable enabling technology” and
affords fragments 2 and 3. Fragment 3 was prepared according
to published procedures4 after minor modification but suffers
disadvantages of low yield and inevitable chromatography
procedures for purification. Fragments 2 and 3 were speculated
to be combined via a C−N bond-formation reaction.5
Disconnection of B gave fragments 4 and 5, which could be
coupled at a late convergent stage of the synthesis. Fragment 4
is commercially unavailable in large amounts, and its synthesis
involves tedious workup. As shown in the disconnection of C,
fragments 8 and 9 could function in oxazolidinone ring
construction in the final step to yield 1. Despite its linearity,
disconnection of C is better than that of B because of the
commercial availability of compound 8.
We critically reviewed all three routes prior to making a final
decision and found that all of these routes involve the common
key building block 2, 2-(1H-pyrazol-4-yl)pyridine (Figure 2).
This building block could be facilely obtained through the
reaction of 2-(pyridin-2-yl)malonaldehyde with hydrazine in
gram scale according to the literature.6 However, 2-(pyridin-2-
yl)malonaldehyde is commercially unavailable and difficult to
Figure 1. Structures of linezolid and compound 1.
treat serious Gram-positive infections in clinics since it was
approved by the FDA in 2000.1 A new member of the
oxazolidinone antimicrobial candidates 1 (Figure 1) was found
to be more active in vitro and in vivo and had a pharmacokinetic
profile comparable to that of linezolid (with an apparent
absolute oral bioavailability of 100%). Given these reasons, 1
was selected for preclinical and clinical studies.2 Compound 1
was found through a medicinal chemistry project and required
modest amounts of material for in vitro and in vivo
pharmacological assessment and preliminary toxicity study. As
biological results evolved, larger amounts were necessary to
support the next phase of preclinical studies, and a synthetic
route capable of delivering multikilogram quantities was
investigated.
The original medicinal chemistry synthetic routes2 were
certainly able to deliver gram-scale products to support studies
of SARs, preliminary toxicities, and physicochemical properties.
However, the unstable intermediates and ineffective purification
strategies made their application for a multikilogram campaign
unpractical. Thus, reinvestigation of possible routes for the final
Received: January 26, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 2. Retrosynthetic strategy of compound 1.
Figure 3. Three synthetic approaches of key intermediate 2.
Scheme 1. Summary route to key building block 14 and proposed formation of impurity
produce on a large scale via a green procedure (see Supporting
Information), which led us to investigate novel scalable
methods to prepare compound 2. This study describes our
successful strategy for constructing key building block 2 from
commodity components and an improved route to the key
intermediate 7 in strategies B and C. The method considers the
functional group addition (FGA) approach and adjusts the
sequence of the reaction steps with the aim of addressing safety
issues and tedious separation procedures for compound 2. The
remarkably improved route allowed the successful transition
from gram-scale laboratory preparation to multikilogram
production of 1 in a kilo laboratory.
RESULTS AND DISCUSSION
■
Figure 3 shows that key building block 2 could be produced
using three approaches: (i) cyclization reaction of equivalent of
2-(pyridin-2-yl)malonaldehyde, 2-(1,3-dinitropropan-2-yl)-
pyridine 11 with hydrazine; (ii) Suzuki−Miyaura reaction by
C−C coupling of pyridin-2-ylboronic acid 12 with 4-bromo-
1H-pyrazole or 4-iodo-1H-pyrazole 13; and (iii) the FGA
approach, i.e., addition of one or two amino groups on the
pyrazole ring of compound 2, resulting in different raw
materials.
The first approach involved the discrete synthesis of the
precursor 2-(1,3-dinitropropan-2-yl)pyridine 11 according to
method reported by Roberto7 with minor modifications; this
method used 1,3-dinitropropane as a malondialdehyde
B
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
synthetic equivalent to react with hydrazine and construct a
pyrazole ring in refluxing ethanol.8 However, the application of
the reaction gave only trace amounts of 2. In the second
coupling approach, the hydrogen in the N-1 position of
compound 13 should be substituted with an intact protecting
group before being subjected to the basic Suzuki−Miyaura
reaction conditions. These conditions involve a palladium
catalyst, which is vulnerable to poisoning by nucleophilic attack.
In addition, expenses associated with pyridin-2-ylboronic acid
12 also impede the implementation of this approach to produce
compound 2 on a pilot scale.
Considering these issues, we investigated the FGA approach,
which presented us with two options to access compound 2
after adding one or two amino groups to its pyrazole ring. First,
for symmetry, two amino functionalities can be added in the
pyrazole ring of compound 2 and reversely derived to 4-
(pyridin-2-yl)-1H-pyrazole-3,5-diamine 14a. This reverse step
can be carried out by condensation reaction9 of 2-(pyridin-2-
yl)malononitrile with hydrazine. A key point here is the
efficient synthesis of 2-(pyridin-2-yl)malononitrile. In the
laboratory, production from 2-bromopyridine and malononi-
trile in suitable conditions is not difficult.10 Nevertheless, this
approach has several disadvantages, such as an expensive rare
metal catalyst, high temperature, and unacceptable yields for
further scale-up. In the second option, 4-(pyridin-2-yl)-1H-
pyrazol-5-amine 14, with one amino group addition, can be
reversely derived to 3-(dimethylamino)-2-(pyridin-2-yl)-
acrylonitrile. Its laboratory gram-scale synthesis can also be
easily accomplished from 2-(pyridin-2-yl)acetonitrile 17 in a
two-step sequence (Scheme 1) with moderate yield by
microwave irradiation11 or normal warm conditions. This
synthetic sequence is more suitable for further optimization and
scale-up because of its significant benefits in workups and
acceptable yields as well as use of inexpensive starting materials
and palladium catalyst. Given this analysis, precursor 14 was
selected for further development.
equiv of hydrobromic acid to give 14 with 78% yield. When an
appropriate catalyst was applied, the protocol allowed the level
of starting material 17 to be lowered to 5%, and the chemical
yield was comparable with that reported. Several acids in
different equiv ratios to the substrate were reported as catalysts
that resulted in good reaction profiles.13 However, this study
determined a protocol maintaining a high level of atom
efficiency and low corrosive effect to the environment. Several
acids were screened to reduce corrosive effects on the
equipment or environment. A laboratory improvement
program was performed to determine whether the equivalents
of acid and hydrazine used in the reaction could be reduced.
Either a weak or strong acid, such as acetic, formic, or
hydrochloric, resulted in high conversion (>90% HPLC peak
area) to the desired 14 and low levels of 2-(pyridin-2-
yl)acetonitrile 17 (<5%). Application of p-toluenesulfonic
acid resulted in significantly decreased conversion. Moreover,
crude 14 was obtained in ∼60% chemical yield, which mostly
associated with 2-(pyridin-2-yl)acetonitrile 17 (∼10%) back-
ward and unknown impurities (∼5%). Acetic acid was used for
the process because formic and hydrochloric acids will result in
more corrosive problems than acetic acid. Upon optimization,
the second improvement was quickly observed. Only 1 equiv of
acid and hydrazine for 12 h resulted in high conversion and a
low level of impurities (Table 1).
Table 1. Product contents of crude intermediate 14
catalyzed with different equivalents of acids
HPLC area %
acid
formic acid
equiv
14
17
4
90.1
93.2
82.2
92.8
93.6
87.0
3.1
4.7
HCl
4
p-toluenesulfonic acid
acetic acid
4
12.4
3.5
4
acetic acid
1
3.6
As depicted in Scheme 1, intermediate 3-(dimethylamino)-2-
(pyridin-2-yl)acrylonitrile 16 was prepared by condensation of
2-(pyridin-2-yl)acetonitrile 17 and DMF−DMA without
solvent and microwave irradiation as required in published
procedures.12 A small amount of fluorescent byproduct 19
formed in the first step was found by TLC under a UV lamp.
This byproduct would convert back to starting material 17 and
desired compound 14 through compound 20 (not isolated) by
reacting with hydrazine in the second step. The reformed 2-
(pyridin-2-yl)acetonitrile 17 can be easily recycled by
suspending the compound 14 in water, filtering workup, and
concentrating, thereby obtaining filtrates under reduced
pressure. The chemical structure of fluorescent byproduct 19
acetic acid
0.1
7.3
Reappearance of starting material 17 could not be avoided in
this process because it arose from both reactions of fluorescent
byproduct 19 with hydrazine and the self-dissociation reaction
of 18. However, the level of 17 was successfully minimized
without downstream difficulties. The sparingly soluble product
14 was purified by making a slurry of the crude product in
water and filtering. Trace amounts of more soluble compound
17 were effectively partitioned into the water phase during
workup such that the final filtered cake containing compound
14 was in relatively high purity (∼99% HPLC area %). While
this process was telescoped, a two-step sequence was
established for the preparation of up to 2 kg batches of 14.
In accordance with the planned synthetic strategy, with an
efficient and cost-effective preparation of 14, we moved forward
to the key intermediate 7 by direct deamination of compound
14 to obtain compound 2, which was followed by C−N
coupling with 1,2-difluoro-4-nitrobenzene. Alternatively, the
fluorinated benzene ring was introduced to compound 14 to
make the regioisomeric mixture 21, which was followed by a
deamination step to produce the key intermediate 7. The
differences between these two measures lay not only in the
reversed synthetic sequence but in significant effects on the
safety issues and separation workups as described below.
Initially, amine pyrazole 14 was directly subjected to the
deamination step (Scheme 2). The use of hypophosphorous
1
was isolated and confirmed by H NMR and MS spectra.
Because fluorescent byproduct 19 cannot be conveniently
removed in the first step and would partially convert to the
desired product in the second step, purification of the crude
intermediate 16 was necessary after conventional removal of
excess reactant DMF−DMA in reduced pressure.
The second step for the synthesis of compound 14 gave us
problems because of the absence of an acid catalyst in refluxing
ethanol during our first kilogram campaign. Compound 14 was
obtained with a lower chemical yield of 34% with a huge
backward formation of 2-(pyridin-2-yl)acetonitrile 17. Some
studies suggest that moderate to good yields (50−80%) can be
achieved if the appropriate acid catalyst is utilized. McCall14
used acrylonitrile to react with hydrazine in the presence of 4
C
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 2. Three conditions for the deamination of
compound 14
Figure 4. Chemical structure of byproduct 22.
Thus, its electrophilic reactivity was enhanced and resulted in
nonselective electrophilic attack of the secondary as well as the
primary amine functionality in compound 14. This unaccept-
able nonselectivity triggered screening for a less polar solvent
with suitable base to synthesize compound 7 with high purity.
Screening of conditions was performed using microscale
high-throughput experimentation strategy. As we expected, the
use of less polar solvents gave no impurity 22, which was
detected in the reaction when DMSO was used as solvent.
Initial attempts to identify the reaction temperature were
performed using DMF solvents and K2CO3 as base. Reactions
under temperatures of 25 and 45 °C were not ideal with low to
moderate conversion of compound 10; conversion was as high
as 95% at 75 °C. Thus, the following development work was
carried out at 75 °C. This high conversion was not reproduced
when DMF was used as solvent in the first batch of screening,
which did not progress beyond 70% conversion. A byproduct
23 was also produced, the chemical structure of which was
elucidated by NMR and MS spectra (see Supporting
Information). This byproduct 23 arose from the coupling
reaction of 1,2-difluoro-4-nitrobenzene 10 with dimethylamine,
which was supposed to arise from hydrolysis of DMF in the
presence of moisture at 75 °C (Scheme 4). To address this
acid is one of the conventional methods in the literature for this
deamination in the presence of sodium nitrite.14 The
application of this acid resulted in a modest chemical yield
(∼60%), but the reaction suffered from tedious workup. This
problem was mostly due to the very viscous and sticky reaction
mixture, which could be resolved by using more volumes of
water as solvent but would significantly decrease the process
batch size. As depicted in Scheme 2, two more conditions for
the deamination were specified.15 Upon scaling up to 20 g of
starting material 14, these two conditions were used separately.
Safety issues of a delayed sharp exothermic event and
occurrence of N2 release on this laboratory scale were observed,
which are expected to present potential hazards during
kilogram scale-up. Using isoamyl nitrite as the deamination
reagent would produce more byproducts, which increases the
isolation difficulty; however, we constructed the key building
block up to 1 kg batches by this method. The preparation of 2
by the direction deamination method was still unsatisfactory for
even larger-scale batches because of important safety issues or
isolation difficulties.
Consequently, the synthesis of key intermediate 7 by
preparing the planned 2 followed by coupling with 1,2-
difluoro-4-nitrobenzene 10 was not an ideal strategy. To
address the aforementioned issues, we wondered whether or
not reversion of the two-step synthetic sequence would provide
us a more attractive alternative leading to the key intermediate
7 (Scheme 3).
Scheme 4. Proposed route to the formation of impurity 23
Scheme 3. Synthetic routes of compound 7
issue resulting from water, N2 atmosphere was charged in our
subsequent development work. In the second batch of
screening, the reaction, which was charged with N2 using
DMF as solvent, provided high conversion with no byproduct
23 detected.
Figure 5 shows that the efficiency of formation of 21 was
found to be highly dependent on the polarity of solvent as well
as the strength of the employed base. Experimental results
suggested that the solvents with high polarity gave better
conversions. For example, reactions in dioxane or acetone
showed slow conversion rates and stalled at less than 90%
conversions, whereas reactions in ACN or DMF demonstrated
high conversion rates of >95% for 5 h at 75 °C when Cs2CO3
was used as the base for deprotonation of 14. Similar trends
were observed for the strength of the employed bases, i.e. the
stronger base produced higher conversions compared with its
weaker counterparts. Notably, among these screened bases,
potassium and cesium were both better counterions for
deprotonation than sodium. However, when K2CO3 was used
as base, the reaction conversion in ACN significantly decreased
(95.6% to 53.1%), whereas DMF resulted in a good conversion
In the gram-scale synthesis of regioisomeric mixture 21, 1.0
equiv of 1,2-difluoro-4-nitrobenzene 10 and K2CO3/DMSO at
various temperatures (25, 45, and 75 °C) were used arbitrarily.
The reaction progress was monitored by HPLC. The different
temperatures gave similar reaction profiles, which were high
conversion but low HPLC purity with a level of 10−20%
byproduct of 22 (Figure 4). The formation of 22 was mostly
attributed to a strong polarization effect of DMSO, which
activated the C−F bond in 1,2-difluoro-4-nitrobenzene 10.
D
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 5. Different reaction conditions for deamination to
make key intermediate 7
workup was still involved, where large amounts of solvent and
repetitions of the extraction procedure were needed to warrant
major isolation of the products from the aqueous phase. In
condition 3, an altered feeding manner guaranteed a good
safety with mild nitrogen release and no delayed sharp
exothermic event with an adiabatic temperature rise value of
<17.5 °C after the completion of feeding the amine compound
21. In previous attempts, amine substrate 21 was charged in the
reactor and dissolved in DMF. Then, isoamyl nitrite was added
in one portion (slow addition made the product quite impure),
which resulted in uncontrollable nitrogen release and increased
reaction temperature. When the anhydrous DMF was heated to
55 °C, isoamyl nitrite was poured into the reactor in one
portion with vigorous agitation. After heating, the solid amine
21 was then added in batches. The feeding speed was
controlled to ensure a mild gaseous release, and a period of
40 min was appropriate on a scale of 2 kg of solid amine 21.
The obtained product was of high purity, and this reaction
feeding manner was suitable for scalable synthesis. The
aforementioned optimized feeding manner was successfully
applied in the preparation of key intermediate 7 up to 2.1 kg in
one batch.
Figure 5. HPLC results of the C−N coupling reaction in different
conditions at 75 °C.
of 98.4%. Meanwhile, considering the cost and hygroscopic
property of Cs2CO3, K2CO3 was selected as the base promoter.
More repetitions of this reaction were performed in gram
scales. Optimal conditions were found and confirmed by using
1.2 equiv of 1,2-difluoro-4-nitrobenzene 10 and 1.5 equiv of
potassium carbonate in DMF at 75 °C. Using these conditions,
a kilogram batch was performed, and compound 21 yielded
96% without contaminants. The product was notably a mixture
of 3-amino and 5-amino regioisomers in a ratio of about 3:1 by
1H NMR (Figure 6). Separation of the isomeric mixture was
not necessary because the amino groups would be removed in
later steps.
All reasonable endeavors for the key intermediate 7 exhibited
good results. The following steps (Scheme 6) were
implemented according to conventional reaction conditions
although some major improvements had also been made during
our development.
Reduction of 7 with iron powder in the presence of
concentrated hydrochloric acid at reflux in EtOH cleanly
afforded 6 in 67% yield. This reaction condition was not
suitable for industry because of its significant environmental
effect from the iron slurry. The alternative for preparing 6
involved reduction of 7 with FeCl3·6H2O and hydrazine in the
presence of activated carbon at reflux. Notably, the safety issue
of potential pressure hazard that was caused by evolution of N2
was minimized by the slow hydrazine feeding rate and favorable
agitation speed. The reaction mixture was warmly filtered
through a short pad of silica to remove FeCl3 and activated
carbon powder, with the filtering cake washed with hot ethanol.
After evaporation workup, the residue was dissolved in
anhydrous acetone, with a small amount of hydrazine
hydrochloride remaining at the bottom, which could be
removed by decantation. The acetone solution of compound
6 was used directly in the following N-acylation reaction to give
intermediate 9.
Figure 6. Component ratio determination of regioisomers of 21 by 1H
NMR.
Thus, the key intermediate 2-(1-(2-fluoro-4-nitrophenyl)-
1H-pyrazol-4-yl) pyridine 7 could be synthesized by deami-
nation of the corresponding regionisomeric mixture 21 as
depicted in Scheme 5.
To find an effective and efficient condition suitable for a safer
deamination process of 21, conditions used for deamination of
14 to give 2 were again investigated in this step. Condition 1
resulted in a similar reaction profile stated in the reaction of
deamination of 14, which also suffered from a delayed sharp
exothermic event and nitrogen release. The application of
condition 2 on this reaction resulted in the absence of sticky
and viscous materials. However, a tedious and time-consuming
N-acylation of 6 with benzyl chloroformate (Cbz-Cl) in the
presence of K2CO3 at 25 °C in DCM afforded 9 in a yield
around 90%. Although this condition resulted in a high yield,
extraction workup was needed, and that suffered from a
significant emulsification problem which was time- and solvent-
consuming. The product was also contaminated with colored
E
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 6. Final process synthetic route of compound 1
impurities, which were difficult to remove by crystallization
regardless of solvent. When the resultant colored material was
directly carried into the next step, it presented the same
downstream difficulties. Thus, after careful development, the
conditions of the Schotten−Baumann acylation modification
were finally used for the synthesis of 9. These conditions with
dilute aqueous K2CO3 in acetone required 2.0 equiv of Cbz-Cl.
When the reaction was complete, a concise solid−liquid
separation removed most of the impurities, including a small
amount of unreacted 7, (carried over from the previous step),
amine 6, and inorganic salts. The filtered cake was resuspended
in water to remove the trace amounts of attached inorganic
salts. Crystallization from 50% ethanol allowed the removal of
excess Cbz-Cl, and the isolated product was consistently pale
white with no colored impurity. After drying in an oven to a
constant weight at 80 °C, the water content, which would
consume lithium tert-butoxide in the subsequent chemical
reaction, was 0.8% as determined by Karl Fischer titration. This
content met our practical demand.
Finally, application of the literature conditions to convert
compound 9 gave expected yields and high conversions to the
oxazolidinone.16 Significant impurities were not observed as
reported by studies,17 and unknown minor impurities were
removed by crystallization. We attempted various solvent
systems, but with minimal success, except for the EtOH and
H2O (3:1) mixture, which removed most of the colored
impurities and lithium salt. Final active pharmaceutical
ingredient (API) 1 with desired polymorph was obtained by
straightforward solid−liquid separation with a high isolated
chemical yield of 82%. This material was of pharmaceutically
acceptable purity (99.9 HPLC area %) and met all the required
specifications. Thus, the final synthetic route for API was
established as shown in Scheme 6.
CONCLUSION
■
In summary, the key building block 14 was successfully
prepared using a catalyst of 1 equiv of acetic acid. This
compound was produced on a large scale and high yield after
the proposed mechanism was telescoped. The key intermediate
2-(1-(2-fluoro-4-nitrophenyl)-1H-pyrazol-4-yl)pyridine, 7, lead-
ing to the API was made by reversing the two-step reaction
sequence from the key building block 14, thereby providing
safety guarantees and facile solid−liquid separation workups.
The merits of the environmentally benign synthetic strategy of
1 are highlighted by its acceptable cost because of its readily
available starting materials, only five isolation procedures for
seven sequential transformations, and no laborious separation
work-ups involved throughout the route. This study presents a
process that allows rapid delivery of a multikilogram quantity of
API with pharmaceutically acceptable purity and an overall
yield of 27.6%.
EXPERIMENTAL SECTION
■
All solvents and reagents were purchased from the suppliers
and used without further purification. H NMR spectra were
1
recorded on a Bruker Avance (Varian Unity Inova) 400 MHz
spectrometer in CDCl3 or DMSO-d6 with Me4Si (TMS) as an
internal standard. Mass spectra were recorded on a Waters Q-
TOF Premier mass spectrometer. HPLC analyses were
performed on a Waters Alliance system (Waters Corp., Milford,
MA, U.S.A.), which was equipped with a performance PLUS
inline degasser along with an autosampler and a 2998
photodiode array detector. LC/MS analyses were carried out
using a Shimadzu LC-20AD interfaced to an Agilent AB SCIEX
3200 Q TRAP mass spectrometer equipped with an electro-
spray ionization (ESI) source. Separation was achieved using an
F
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
XBridge BEH C18 XP column (75 mm × 2.1 mm, 2.5 μm) from
Waters. This column has a mobile phase consisting of methanol
and an aqueous solution of formic acid (0.01%, v/v) at a flow
rate of 0.3 mL/min. The MS analysis was performed in Q1MS
mode.
Hz, 0.3 H), 6.89 (s, 0.6 H), 6.49 (s, 2 H); ESI-MS m/z 300.0
(M + H+).
2-(1-(2-Fluoro-4-nitrophenyl)-1H-pyrazol-4-yl)-
pyridine (7). A 50-L reactor was charged with DMF (30 L)
and heated to 55 °C, and 3-methyl-1-nitroso-oxybutane (2.43
L, 18 mol) was added in one portion. The solid 21 (2.7 kg, 9
mol) was then added to the mixture in batches to stabilize the
nitrogen release rate (on this scale, 40 min was required to
complete the addition). During addition of 21, the temperature
was maintained below 75 °C (the addition was exothermic).
The reaction mixture was slowly cooled to 55 °C and stirred for
2 h. When HPLC analysis indicated complete conversion, the
water (90 L) was charged with stirring. The resulting slurry was
cooled to 25 °C, allowed to stand for more than 1 h, and
filtered. The filtered cake was washed with water (3 L × 3) and
dried to constant weight in an oven at 80 °C to give 7 (2.1 kg,
yield 81.9%) as a brown solid. The HPLC purity was over
(E)-3-(Dimethylamino)-2-(pyridin-2-yl)acrylonitrile
(16). A reactor was charged with 2-pyridylacetonitrile (17, 2 kg,
16.93 mol) and DMF−DMA (3.31 L, 24.89 mol). The reaction
mixture was heated to reflux for 3 h. The progress of the
reaction was monitored by HPLC for the complete absence of
starting material 17. When the reaction was completed (conv.
>99%), the solution was evaporated to dryness under reduced
pressure at 60 °C. The solid residue was used without further
purification in the next step (100% yield of 17 was assumed for
1
calculations). H NMR (400 MHz, CDCl3) δ 8.34 (d, J = 4.4
Hz, 1 H), 8.06 (s, 1 H), 7.58 (t, J = 7.7 Hz, 1 H), 7.39 (d, J =
8.1 Hz, 1 H), 6.98−6.91 (m, 1 H), 3.31 (s, 6 H); ESI-MS m/z
174.4 (M + H+).
1
98.8%. H NMR (400 MHz, DMSO-d6): δ 9.09 (s, 1 H), 8.69
(d, J = 5.2 Hz, 1 H), 8.62 (s, 1 H), 8.50 (dd, J = 11.6 Hz, J = 2.4
Hz, 1H), 8.30−8.21 (m, 2 H), 7.50−7.46 (m, 2 H), 7.11 (d, J =
8.8 Hz, 1 H). 13C NMR (100 MHz, DMSO-d6) δ 150.61,
150.02, 145.99, 141.33, 137.52, 133.12, 130.25, 126.45, 124.59,
122.61, 121.33, 120.71, 114.13, 111.37; ESI-MS m/z 285.13 (M
+ H+).
4-(Pyridin-2-yl)-1H-pyrazol-3-amine (14). The afore-
mentioned residue (2.33 kg, 13.45 mol) was dissolved in
EtOH (11.5 L) in a 50 L reactor. The mixture was heated to
reflux for 30 min and cooled to 25 °C. Then, acetic acid (759
mL, 13.25 mol) and hydrazine hydrate (805 mL, 16.6 mmol)
were added successively into the reaction mixture with stirring.
The mixture was heated to reflux for 12 h. The progress of the
reaction was monitored by HPLC for the complete absence of
16. When the reaction was completed, the mixture was
evaporated to form a thick slurry. The resultant slurry was then
transferred into a separate vessel, and water (2 L) was added
under constant stirring. The mixture was filtered under vacuum.
The solid collected was then washed with water (1 L × 3) and
air-dried on the filter under suction and then dried to a
constant weight in an oven at 60 °C. The result was a solid
beige 14 (1.68 kg, two steps, 63%), which was used in the next
step without further processing. The HPLC purity was over
3-Fluoro-4-(4-(pyridin-2-yl)-1H-pyrazol-1-yl)aniline
(6). Compound 7 (2.0 kg, 7.04 mol), FeCl3 (400 g), EtOH (20
L), and activated carbon were added consecutively to a reactor.
The mixture was heated to reflux and stirred vigorously. To this
solution was dropwise added hydrazine hydrate (2.82 kg, 56.28
mol; on this scale, 80 min was required to complete the
addition) with stirring while maintaining reflux. The progress of
the reaction was monitored by HPLC for the complete absence
of 7. When coupling was completed as indicated by HPLC
(conv. >99%), iron salts and activated carbon powder were
removed by filtration through a pad prepared from silica gel (1
kg). The pad was then rinsed with ethanol (1 L), and the
filtered cake was washed with ethanol (2 L × 2). The filtrate
was then concentrated (partial vacuum and jacket temperature
<60 °C) to dryness to give crude 6, a pale-brown solid
containing hydrazine hydrochloride. This obtained product was
directly used in the next step without further purification (100%
1
98.5%. H NMR (400 MHz, CDCl3) δ 11.78 (br s, 1 H), 8.43
(d, J = 4.8 Hz, 1 H), 7.91 (br s, 1 H), 7.66 (t, J = 7.6 Hz, 1 H),
7.55 (d, J = 8 Hz, 1 H), 7.00 (t, J = 5.6 Hz, 1 H), 5.96 (br s, 2
H); ESI-MS m/z 161.4 (M + H+).
Regioisomeric Mixture of 1-(2-Fluoro-4-nitrophenyl)-
4-(pyridin-2-yl)-1H-pyrazol-3(5)-amine (21). A reactor was
charged with 14 (1.58 kg, 9.86 mol), K2CO3 (2.1 kg, 14.79
mol), and DMF (32 L), and the resulting mixture was stirred
for 30 min at 25 °C. Then 1,2-difluoro-4-nitrobenzene (10,
1.33 kg, 11.83 mol) was added into the reaction mixture in one
portion. After the reactor was degassed under vacuum, the
atmosphere was exchanged with nitrogen. The mixture was
heated to 75 °C and stirred for about 5 h. The progress of the
reaction was monitored by HPLC for the complete absence of
14. When coupling was completed as indicated by HPLC (conv
>99%), water (64 L) was added into the slurry with stirring.
Then, the resulting slurry was stirred for 10 min, cooled to 25
°C, and filtered. The filtered cake was washed with deionized
water to remove base substance and dried to constant weight in
an oven at 80 °C to give 21 (2.83 kg, 96% yield) as a yellow
1
yield of 7 was assumed for calculations). H NMR (400 MHz,
DMSO-d6): δ 9.89 (s, 2 H), 9.48 (s, 1 H), 8.88 (s, 1 H), 8.73
(d, J = 6 Hz, 1 H), 8.58−8.54 (m, 1 H), 8.46 (d, J = 8.4 Hz, 1
H), 7.84 (t, J = 6.8 Hz, 1 H), 7.66 (t, J = 8.4 Hz, 1 H), 7.04 (d, J
= 12.8 Hz, 1 H), 6.95 (d, J = 8.8 Hz, 1 H); ESI-MS m/z 255.16
(M + H+).
Benzyl 3-fluoro-4-(4-(pyridin-2-yl)-1H-pyrazol-1-yl)-
phenylcarbamate (9). The previously mentioned crude 6
(1.79 kg, 7.04 mol) was dissolved in acetone (11 L) with
undissolved hydrazine hydrochloride remaining at the bottom.
The supernatant was decanted to a reactor, and the hydrazine
hydrochloride was discarded. The reactor containing a solution
of 6 in acetone (11 L) was charged with potassium carbonate
solution (1.46 kg K2CO3/7 L H2O) with agitation. The
resulting mixture was cooled to 5 °C and stirred for 1 h. To this
solution was dropwise added Cbz-Cl (1.68 L, 13.9 mmol) with
stirring, and the temperature of the reaction mixture was
maintained under 15 °C. When the addition was completed
(on this scale, 80 min was required to complete the addition),
the temperature of the reaction was cooled to 25 °C and stirred
for 12 h. The HPLC analysis of a reaction aliquot indicated
>99% conversion of raw material 6. The resulting slurry was
1
solid. The HPLC purity was over 97.9%. H NMR (400 MHz,
CDCl3) δ 8.81 (s, 1H), 8.56 (d, J = 4.8 Hz, 1 H), 8.48 (d, J =
4.8 Hz, 0.3 H), 8.41 (d, J = 10.4 Hz, 0.3 H), 8.37 (d, J = 12.8
Hz, 1 H), 8.26 (s, 0.3 H), 8.21 (t, J = 10.0 Hz, 1 H), 8.09 (t, J =
8.8 Hz, 1 H), 7.93 (D, J = 8.0 Hz, 1 H), 7.88 (d, J = 8.0 Hz, 0.3
H), 7.81 (t, J = 7.6 Hz, 1 H), 7.75 (t, J = 8.0 Hz, 0.3 H), 7.67
(d, J = 8.0 Hz, 0.3 H), 7.23 (t, J = 6.4 Hz, 1 H), 7.07 (t, J = 6.4
G
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Author Contributions
filtered, and the filtered cake was washed with water to remove
the base substance. The obtained solid was suspended in 50%
EtOH (5 L) and heated to reflux for 1 h. The resulting slurry
was cooled to 20 °C, filtered, and dried in an oven to a constant
weight at 45 °C. The result was a solid, pale-white compound 9
(1.83 kg, two steps, 68%). The HPLC purity was more than
§(T.Y. and J.-X.C.) Both authors contributed equally to this
paper.
Notes
The authors declare no competing financial interest.
1
99.0%. H NMR (400 MHz, CDCl3): δ 8.60 (d, J = 4.8 Hz, 1
ACKNOWLEDGMENTS
H), 8.48 (d, J = 2 Hz, 1 H), 8.21 (s, 1 H), 7.82 (t, J = 8.8 Hz, 1
H), 7.72−7.68 (m, 1 H), 7.64 (d, J = 12.8 Hz, 1 H), 7.54 (d, J =
8 Hz, 1 H), 7.43−7.36 (m, 4 H), 7.17−7.14 (m, 1 H), 7.09 (dd,
J = 8.8 Hz, J = 1.2 Hz, 1 H), 6.86 (s, 1 H), 5.23(s, 1 H); 13C
NMR (100 MHz, CDCl3) δ: 154.55, 153.08, 151.37, 149.68,
138.96, 138.04, 137.96, 136.75, 135.69, 129.06, 129.00, 128.67,
128.51, 128.37, 124.92, 124.62, 123.48, 121.54, 119.92, 114.46,
107.00, 67.36; ESI-MS m/z 389.19 (M + H+).
■
The authors greatly appreciate financial support from the
National Major Program of China during the 12th Five-Year
Plan Period (Grant No. 2012ZX09103-101-036) and the New-
Century Excellent Talent Fund by the Ministry of Education
(Grant No. NCET-13-0388).
REFERENCES
■
(S)-N-((3-(3-Fluoro-4-(4-(pyridin-2-yl)-1H-pyrazol-1-
yl)phenyl)-2-oxo-oxazolidin-5-yl)methyl)acetamide (1).
In a 50-L reactor, 9 (1.8 kg, 4.64 mol) and 8 (1.79 kg, 9.27
mol) were dissolved in THF (12.6 L) at −5 °C. The reaction
mixture was degassed by purging with N2. Then, methanol (375
mL, 9.27 mol) was added to the mixture under N2 atmosphere.
After stirring for about 10 min at −5 °C, lithium tert-amylate
(1.11 kg, 13.91 mol) was added to the mixture in one portion
with an exotherm from −5 to 17 °C. The resulting solution was
cooled to −5 °C, yielding a thick slurry, and stirred for about 1
h and stirred again at 25 °C for about 15 h. The slurry was
cooled to 10 °C. The reaction was quenched by adding acetic
acid (525 mL, 9.27 mol) in one portion and stirred for 30 min.
The reaction mixture was then evaporated to dryness at 30 °C.
The solid residue was allowed to soak for 3 h in water (30 L),
stirred, filtered under reduced pressure, and washed with water
(10 L × 3). The solid filtered cake was suspended in ethyl
acetate (10 L). The resulting suspension was heated to reflux
for 2 h, cooled to 25 °C, and filtered under reduced pressure.
The collected solid was resuspended in a mixture of EtOH and
water (6 L/2 L) and heated to reflux for 2 h. The slurry was
cooled to 25 °C, filtered under reduced pressure, and washed
with EtOH (3 L × 2). The filtered cake was dried in an oven to
a constant weight at 45 °C. The final product was an off-white
solid 1 (1.5 kg, isolated yield of 82%). The HPLC purity was
over 99.9%. 1H NMR (400 MHz, CDCl3): δ 8.61 (d, J = 4 Hz,
1 H), 8.52 (d, J = 6.8 Hz, 2 H), 8.22 (s, 1 H), 7.94 (t, J = 8.8
Hz, 1 H), 7.77−7.69 (m, 2 H), 7.55 (d, J = 8 Hz, 1 H), 7.27−
7.26 (m, 1 H), 7.18−7.15 (m, 1 H), 6.06 (t, J = 6 Hz, 1 H),
4.86−4.80 (m, 1 H), 4.11 (t, J = 9.2 Hz, 1 H), 3.86−3.82 (m, 1
H), 3.78−3.62 (m, 2 H), 2.04 (s, 3 H); 13C NMR (DMSO-d6):
δ 170.51, 154.47, 152.94, 151.26, 149.94, 139.70, 139.15,
137.43, 129.96, 125.61, 125.19, 123.42, 122.19, 120.38, 114.52,
106.68, 72.29, 47.70, 41.84, 22.91; ESI-MS m/z 418.08 (M +
Na+).
(1) (a) Brickner, S. J.; Hutchinson, D. K.; Barbachyn, M. R.;
Manninen, P. R.; Ulanowicz, D. A.; Garmon, S. A.; Grega, K. C.;
Hendges, S. K.; Toops, D. S.; Ford, C. W.; Zurenko, G. E. J. Med.
Chem. 1996, 39, 673−679. (b) Barbachyn, M. R.; Ford, C. W. Angew.
Chem., Int. Ed. 2003, 42, 2010−2023.
(2) (a) Gong, C. Y.; Yang, T.; Yang, X. Y.; Liu, Y. Y.; Ang, W.; Tang,
J. Y.; Pi, W. Y.; Xiong, L.; Chang, Y.; Ye, W. W.; Wang, Z. L.; Luo, Y.
F.; Zhao, X.; Wei, Y. Q. Nanoscale. 2013, 5, 275−283. (b) Luo, Y. F.;
Wang, Z. L.; Wei, Y. Q.; Geng F. WO/2012/171479, 2012.
(3) Warren. S.; Wyatt. P. Organic Synthesis: The Disconnection
Approach; Wiley: NewYork, 2008.
(4) (a) Genin, M. J.; Allwine, D. A.; Anderson, D. J.; Barbachyn, M.
R.; Emmert, D. E.; Garmon, S. A.; Graber, D. R.; Grega, K. C.; Hester,
J. B.; Hutchinson, D. K.; Morris, J.; Reischer, R. J.; Ford, C. W.;
Zurenko, G. E.; Hamel, J. C.; Schaadt, R. D.; Stapert, D.; Yagi, B. H. J.
Med. Chem. 2000, 43, 953−970. (b) Bank, S.; Marcantonio, R. P. J.
Org. Chem. 1984, 49, 5091−5093.
(5) Kaddouri, H.; Vicente, V.; Ouali, A.; Ouazzani, F.; Taillefer, M.
Angew. Chem., Int. Ed. 2009, 48, 333−336.
(6) Roppe, J.; Smith, N. D.; Huang, D.; Tehrani, L.; Wang, B.;
Anderson, J.; Brodkin, J.; Chung, J.; Jiang, X.; King, C.; Munoz, B.;
Varney, M. A.; Prasit, P.; Cosford, N. D. P. J. Med. Chem. 2004, 47,
4645−4648.
(7) Roberto, B.; Giovanna, B.; Dennis, F.; Alessandro, P. Synthesis.
2004, 12, 1938−1940.
(8) Cabrera, E. F.; Derri Alcdntara, M. P.; Gomez-Sanchez, A.
Tetrahedron Lett. 1988, 29, 6001−6004.
(9) Mccall, J. M.; Kelly, R. C.; Romero, D. L. Pat, 2012, WO/2012/
149157.
(10) Schnyder, A.; Indolese, A. F.; Maetzke, T.; Wenger, J.; Blaser, H.
U. Synlett. 2006, 18, 3167−3168.
(11) Sun, L.; Bera, H.; Chui, W. K. Eur. J. Med. Chem. 2013, 65, 1−
11.
(12) (a) Behbehani, H.; Ibrahim, H. M.; Makhseed, S.; Mahmoud, H.
Eur. J. Med. Chem. 2011, 46, 1813−1820. (b) Behbehani, H.; Ibrahim,
H. M.; Makhseed, S.; Elnagdi, M. H.; Mahmoud, H. Eur. J. Med. Chem.
2012, 52, 51−65. (c) Velankar, A. D.; Quintini, G.; Prabhu, A.; Weber,
A.; Hunaeus, G.; Voland, B.; Wuest, M.; Orjeda, C.; Harel, D.;
Varghese, S.; Gore, V.; Patil, M.; Gayke, D.; Herdemann, M.; Heit, I.;
Zaliani, A. Bioorg. Med. Chem. 2010, 18, 4547−4559.
ASSOCIATED CONTENT
■
(13) (a) Xu, Y.; Brenning, B. G.; Kultgen, S. G.; Liu, X.; Saunders,
M.; Ho, K. WO/2013/013188 2013. (b) Anwar, H. F.; Fleita, D. H.;
Kolshorn, H.; Meier, H.; Elnagdi, M. H. ARKIVOC 2006, 15, 133−
141. (c) Salaheldin, A. M.; Alphy, M. K. J. Heterocycl. Chem. 2008, 45,
307−310.
S
* Supporting Information
Analytical methods for the intermediates and API, synthetic
methods for 2-(pyridin-2-yl)malonaldehyde in laboratory scale
and spectral data of process-related impurities. This material is
available free of charge via the Internet at http://pubs.acs.org.
(14) Kornblum, N. Org. Synth. 1941, 21, 30.
(15) (a) Toto, P.; Chenault, J.; Hakmaoui, A. E.; Akssira, M.;
Guillaumet, G. Synth. Commun. 2008, 38, 674−683. (b) Saenz, J.;
Mitchell, M.; Bahmanyar, S.; Stankovic, N.; Perry, M.; Craig, W. B.;
Kline, B.; Yu, S.; Albizati, K. Org. Process. Res. Dev. 2007, 11, 30−38.
(16) Perrault, W. R.; Pearlman, B. A.; Godrej, D. B.; Jeganathan, A.;
Yamagata, K.; Chen, J. J.; Lu, C. V.; Herrinton, P. M.; Gadwood, R. C.;
AUTHOR INFORMATION
■
Corresponding Author
*E-Mail: luo_youfu@scu.edu.cn
H
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Chan, L.; Lyster, M. A.; Maloney, M. T.; Moeslein, J. A.; Greene, M.
L.; Barbachyn, M. R. Org. Process. Res. Dev. 2003, 7, 533−546.
(17) Lu, C. V.; Chen, J. J.; Perrault, W. R.; Conway, B. G.; Maloney,
M. T.; Wang, Y. Org. Process Res. Dev. 2006, 10, 272−277.
I
dx.doi.org/10.1021/op500030v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX