Enabled Process to Synthesize Monobactam 1 for Early Development

Article abstract of DOI:10.1021/acs.oprd.9b00374

An efficient process to synthesize Monobactam 1 from intermediates 2, 3, and 4 was developed. The process was initially enabled to deliver 500 g of 1 for exploratory toxicity studies and further improved to facilitate kilogram-scale production of the active pharmaceutical ingredient with enhanced quality and throughput. Highlights of the process encompass utilizing 1,1′-carbonyldiimidazole to construct a urea linker with good selectivity, oxidative cleavage to remove a 2,4-dimethoxybenzyl protecting group, two activation methods to form an amide bond, a high yielding reaction to introduce N-sulfonic moiety, and a global deprotection to remove all protecting groups. Practical procedures were developed to isolate intermediates en route by adding a concentrated substrate to antisolvents to obtain filterable amorphous solids with partial purification. Amberchrom CG161M resin was applied not only as a "resin-capture-release" tool to remove the bulk amount of water but also as an effective method to purify 1. Finally, a process isolating 1 pentahydrate (1·5H₂O) as a crystalline solid from the acidic aqueous solution was developed based on zwitterionic crystallization methodology.

Full text of DOI:10.1021/acs.oprd.9b00374

Article
pubs.acs.org/OPRD
Cite This: Org. Process Res. Dev. XXXX, XXX, XXX−XXX
Enabled Process To Synthesize Monobactam 1 for Early
Development
Yong Tao,* Manjinder S. Lall,* David C. Boyles, Susan C. Lilley, Sebastian D. Pattavina,
Robert J. Rafka, Barbara J. Sitter, Andrew Morgan Stewart, III, Jan Szeliga, and Gerald A. Weisenburger
Pfizer Worldwide Research and Development, Eastern Point Road, Groton, Connecticut 06340, United States
ABSTRACT: An efficient process to synthesize Monobactam 1 from intermediates 2, 3, and 4 was developed. The process was
initially enabled to deliver 500 g of 1 for exploratory toxicity studies and further improved to facilitate kilogram-scale production
of the active pharmaceutical ingredient with enhanced quality and throughput. Highlights of the process encompass utilizing
1,1′-carbonyldiimidazole to construct a urea linker with good selectivity, oxidative cleavage to remove a 2,4-dimethoxybenzyl
protecting group, two activation methods to form an amide bond, a high yielding reaction to introduce N-sulfonic moiety, and a
global deprotection to remove all protecting groups. Practical procedures were developed to isolate intermediates en route by
adding a concentrated substrate to antisolvents to obtain filterable amorphous solids with partial purification. Amberchrom
CG161M resin was applied not only as a “resin-capture−release” tool to remove the bulk amount of water but also as an
effective method to purify 1. Finally, a process isolating 1 pentahydrate (1·5H2O) as a crystalline solid from the acidic aqueous
solution was developed based on zwitterionic crystallization methodology.
KEYWORDS: unsymmetrical urea, oxidative DMB cleavage, amide bond formation, global deprotection, Amberchrom CG161M,
resin-capture−release, zwitterionic crystallization
Urea Formation. Among many activating reagents used to
construct an unsymmetrical urea,4 1,1′-carbonyldiimidazole
(CDI) is the most safe and nontoxic one.5 Initial tests revealed
that the reaction of CDI with 4 in dichloromethane (Scheme
2) offered a superior profile to form the carbamoylimidazole
intermediate compared to the reaction with 2. It was also
found that using a free base of 4 in anhydrous conditions was
required to obtain clean and stable carbamoylimidazole 10.
The optimal molar ratio of CDI, 4, and 2 was then identified at
1.5:1:1, respectively, that provided >90% of pure desired urea
7 with low levels of unreacted 2 (<2%) and 4 (<1%), and side
symmetric ureas 8 (∼5%) and 9 (∼2%) (Figure 1). Thus,
neutralizing the mesylate salt of 4 with K3PO4 in water and
dichloromethane, followed by drying the organic layer with
magnesium sulfate, offered an anhydrous solution of free base
4 in dichloromethane (<0.2% wt. of water), which was slowly
added to a solution of CDI (1.5 equiv) in dichloromethane at
20 °C to form 10. A stress test indicated that 10 in the solution
was stable for at least 24 h at 20 °C. Due to the slow reaction
of 10 with amine 2, adding dimethylformamide (DMF, ∼2
vol) and increasing the temperature to 40 °C were needed to
accelerate the formation of urea 7. With no success in
attempting crystallization, precipitation through reverse
addition was applied to isolate amorphous 7 with partial
purification. Thus, after solvent replacement with ethyl acetate,
washes with aqueous citric acid and aqueous NaCl purged
most 2, 4, and DMF. The slow addition of the organic layer to
n-heptane precipitated 7, which was easy to filter. On a 6 kg
scale, 7 (95.9% pure) was obtained in a 94% yield. The main
impurities were 8 (2.5%) and 9 (1.3%).
INTRODUCTION
■
Monobactam 11 is a promising antibacterial candidate with a
unique pyridone-conjugated monocyclic β-lactam structure for
the treatment of life-threatening infections caused by Gram-
negative pathogens with multidrug-resistant strains, including
Pseudomonas aeruginosa, Klebsiella pneumoniae, and Escherichia
coli. To fast-track this candidate for early development, we
were initially requested to prepare 500 g of 1 for exploratory
toxicity studies, followed by additional multikilograms of the
active pharmaceutical ingredient (API) for regulatory toxicity
studies and Phase I clinical trials. As described in the previous
papers,1,2 the original synthesis was unable to meet bulk API
demands due to high step count and extremely poor
throughput (∼0.2% overall yield). Therefore, a more concise
and efficient synthesis to 1 was needed. Previously, we
reported a process to prepare kilogram quantities of
monocyclic β-lactam core (2).2 Herein, we describe the
process development to synthesize 1 from 2, 3 (the “left-hand”
thiazole piece, LHP),3 and 4 (the “right-hand” pyridone piece,
RHP).1
DISCUSSION AND RESULTS
■
Route Selection. The new synthetic sequence was
designed with orthogonal considerations in mind, which
utilized intermediates 2, 3, and 4 with minimal protecting
group manipulation. In recognizing the t-butyl ester of 3 being
relatively labile to acidic conditions and having the new core
(2) bearing the 4-aminomethyl moiety in hand, a logical
approach assembling the target molecule was first constructing
the urea linker between 2 and 4, followed by forming the
amide bond with 3, then introducing the N-sulfonic acid
moiety, and finally removing all protecting groups to deliver
monobactam 1 (Scheme 1).
Received: August 24, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Comparison of Bond Disconnections between Original and New Syntheses toward 1
Scheme 2. Formation of Urea 7 through CDI Activation
Consequently, a two-step deprotection approach that removed
first DMB group through oxidation and then N-Boc group by
acid was utilized. On finding that reaction 7 with ceric
ammonium nitrate in aqueous acetonitrile gave only a 3:1
mixture of desired 11 and side amide 12 (Figure 2),6 oxidative
DMB cleavage with potassium persulfate in the presence of
K2HPO4 in aqueous acetonitrile7 was tried, and a complete
reaction was achieved at 70 °C without forming 12 (Scheme
3). Further studies revealed that using 4 equiv of potassium
persulfate and 4.5 equiv of K2HPO4 to set the initial pH in 6−7
range and keeping the temperature around 70 °C were
essential for both reaction initiation and fast progress. Under
these conditions, the reaction was complete in 3 h when the in
situ yield was measured as 65% and the pH of the reaction
Figure 1. Structures of side symmetric ureas.
Removal of 2,4-dimethoxybenzyl (DMB) and N-Boc
Groups. Treating 7 with strong acids, such as trifluoroacetic
acid (TFA), HCl, MsOH, and TsOH, only cleaved the Boc
group and left the 2,4-dimethoxybenzyl (DMB) group intact.
B
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
methyl tert-butyl ether (MTBE) to precipitate crude 14 with
diatomaceous earth as a well-behaved mixed solid. Further
trituration in MTBE purged more TFA and therefore
improved the stability of the mixed solid, which could be
stored at ambient temperature for a couple of weeks with little
purity loss. Using this technique, 3.1 kg of 14 (86.5% pure,
loaded with 8.2 kg of diatomaceous earth) was isolated in an
81% yield on scale.
Figure 2. Structure of side amide 12.
Formation of Amide 19. With 14 (∼27 wt % loaded with
diatomaceous earth) in hand, we developed two procedures to
construct the amide bond. The first method was through
succinimidyl ester 18,9 a stable white solid obtained in an 85%
yield by treating 3 with N-hydroxysuccinimide and N,N′-
diisopropylcarbodiimide (DIC) in dichloromethane. The
reaction of 18 and 14 at a 1:1 molar ratio in the presence of
DMAP (2.5 equiv) and molecular sieves in acetonitrile at 40
°C for 20 h formed amide 19 with >95% conversion (Scheme
5, Method A). Workup through pH adjustment with aqueous
citric acid, solvent exchange to ethyl acetate, filtration to
remove diatomaceous earth and molecular sieves, partial
concentration, and MgSO4 drying offered a solution of 19 in
ethyl acetate, which was slowly added to n-heptane to
precipitate amorphous crude 19 (81.0% pure) in an 83%
yield. As crystallization attempts were fruitless, silica gel
chromatography was used for the first delivery. Thus, 2.35 kg
of purified amorphous 19 (97.8% pure) was obtained in a 64%
yield on a 3 kg scale by precipitation from ethyl acetate and n-
heptane.
An alternative approach to amide 19 (Scheme 5, Method B),
which provided a better reaction profile than Method A,
entailed partially purifying 14 with basic aluminum oxide and
in situ activating acid 3 with 2-chloro-4,6-dimethoxy-1,3,5-
triazine (CDMT). Thus, treating 14 (as mixed solid with
diatomaceous earth) with N-methylmorpholine (NMM) in
acetonitrile, followed by filtering and switching solvent to
DMF, afforded a neutralized 14 solution in DMF. Subjecting
this solution through a short column of basic aluminum oxide
with a DMF rinse produced a ∼95% pure 14 solution in DMF,
while leaving most side products 15, 16, 17, and other small
unidentified impurities on the column. On the other side, the
in situ activation of 3 (∼0.85 equiv) to 20 was achieved by
reacting with CDMT (0.85 equiv) and NMM (1.7 equiv)10 in
ethyl acetate at 20 °C for 2 h. At that point, the partially
purified 14 solution in DMF was added in for amidation,
which took 16 h at 20 °C to complete. After aqueous washes to
purge DMF and concentration of the organic layer, the slow
addition of the solution in ethyl acetate to n-heptane
precipitated amorphous 19 (96.1% pure) in a 69% yield.
This process, which circumvented the chromatography, was
successfully demonstrated on a 100 g scale.
mixture was read at 6.3. However, extending the reaction time
resulted in yield decrease and pH drop (52% yield and pH 4 in
6 h; 13% yield and pH 2 in 18 h). In contrast, periodically
adding K2HPO4 to maintain pH in 6−7 significantly slowed
down the product degradation, and the in situ yield was 62%
after 18 h. With this knowledge on the process stability, the
scale-up of this oxidative cleavage was set up by mixing 7,
potassium persulfate (4.0 equiv), and K2HPO4 (4.5 equiv) in
2:1 acetonitrile and water. The pH was continuously
monitored during the reaction at 70 °C. When the pH
dropped to 6.3, the aqueous K2HPO4 solution was fed to
maintain the pH between 6.3 and 7.0. As oxygen gas was a
byproduct of the reaction, high nitrogen sweep was applied
over the course to keep the oxygen level <6% in the headspace
to meet our safety requirements (<12%). After the reaction
completion in 2 h, most inorganic precipitates were removed
by filtration and the organic solvent was switched from
acetonitrile to ethyl acetate to crystallize crude 11. Further
trituration in biphasic ethyl acetate and water purged more
inorganic salts and organic impurities, such as aldehyde 13 and
urea 8. Sufficient drying of 11 to residual water <0.5% was
required for the next step. On a 4 kg scale, 11 (96.6% pure)
was obtained in a 61% yield.
The main challenge to deprotect N-Boc from 11 was that
the procedure called for concentrated strong acid, which
caused degradation of product 14 over time. Initial experi-
ments revealed that neat TFA gave a superior reaction profile
at low temperature compared to HCl in dioxane or TFA in
dichloromethane. Reacting with large excess (∼40 equiv) of
TFA in the presence of anisole (2 equiv) as a cation scavenger8
at −10 to −5 °C for 2−4 h was concluded as the optimal
condition, which offered a typical reaction profile with ∼90%
of 14 and ∼10% of the proposed side product 15 (Scheme 4).
Liquid chromatography−mass spectrometry inference indi-
cated that 14 could further react with TFA to form 16, and 15
to 17. Due to the poor stability of 14 in strong acid at elevated
temperature, isolating 14 on scale was initially problematic
with significant product degradation in removing TFA under
vacuum distillation. This challenge was overcome by
implementing a precipitation protocol without heat. Con-
sequently, the reaction solution in TFA was slowly added to a
cooled suspension of diatomaceous earth (220 wt % to 11) in
N-Sulfonation to 21. Converting 19 to 21 with 10 equiv
of sulfur trioxide−DMF complex in DMF11 was fast at ambient
Scheme 3. Balanced Equation of Oxidative DMB Cleavage To Prepare 11
C
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 4. N-Boc Deprotection of 11 and Proposed Side Products
Scheme 5. Methods A and B to Amide 19
Scheme 6. N-Sulfonation to 21
temperature (Scheme 6), and the reaction was complete in less
than 30 min with forming side bis-sulfonated product 22 (2−
15%). An attempt to eliminate the formation of 22 by lowering
the amount of the sulfonating reagent to <8 equiv only resulted
in an incomplete conversion. Fortunately, 22 could be cleanly
converted to 21 in the workup by treating the mixture with
MgSO4 in the presence of a small amount of water.12 Thus,
quenching the reaction to a biphasic mixture of ethyl acetate
and aqueous NaCl, removing DMF by aqueous washes, and
drying the wet organic layer with MgSO4 under vigorous
agitation for 1 h converted 22 to 21. Concentrating the filtrates
and then adding MTBE precipitated amorphous 21 (89.8%
pure) in an 87% yield on a 2.3 kg scale.
Global Deprotection To Target 1. For the final step
removing all protecting groups to form 1 (Scheme 7), two-step
deprotection approaches were initially explored. Approach 1
was able to deprotect N-Boc and t-butyl groups by treating 21
with HCl in dioxane to give intermediate 23, which
surprisingly would not undergo debenzylation under standard
hydrogenation conditions. In contrast, the reversed sequence
(Approach 2) offered a smooth debenzylation under hydro-
genation with Pd/C; however, N-Boc and t-butyl deprotection
D
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 7. Deprotection Strategy to Monobactam 1
under acidic conditions were not clean. On the other hand,
strong Lewis acids, such as BCl3 and BBr3, have been widely
used to cleave O-benzyl, carbamates, and ester bonds1,13 and
would provide an ideal one-step “global deprotection” in our
case if successful. Indeed, treating 21 with excess BCl3 in
dichloromethane at −5−10 °C cleaved all protecting groups
(e.g., t-butyl ester, N-Boc, and two O-benzyl groups) within 1
h. An optimal amount of 10 equiv of BCl3 ensured reaction
completion without significant impact to the product stability
and afforded ∼65% pure 1 with numerous minor side
products, including monobactam ring-opened impurity (25,
Figure 3). Although the product directly precipitated out, the
was unusual, the project team recognized the need for
qualifying impurities at elevated levels as 1 being a lack of
stability and crystallinity at the outset. Thus, little effort was
placed on obtaining crystalline API at that time; instead, a
reverse-phase chromatography was chosen to accomplish the
API delivery with adequate purity. With the siderophore (N-
hydroxypyridone) liberated at this stage, it was important to
use iron-free solvents, reagents, and equipment going-forward
to avoid chelation with iron, resulting in red to pink product 1
discoloration.1a Preparing the feed solution for the chromatog-
raphy was challenging due to the residual boron occluded in
crude 1 that resulted in foaming and decreasing product purity.
The slow addition of crude 1 to a mixture of DMF and 1 M
aqueous ammonium bicarbonate solution, followed by
filtration, gave a feed mixture (pH ∼4), which was subjected
to a Kromasil C-18 column (9 kg) for purification using
acetonitrile and water (w/0.2% formic acid) as eluent. The
product fractions were concentrated to remove acetonitrile
under 30 °C. To minimize the degradation of 1, a “resin-
capture−release” technique using Amberchrom CG161M
resin15 was applied at this point to remove the bulk amount
of water without heating.16 Thus, the aqueous residue was
passed through the resin column with a water wash to capture
product 1 on the resin, which was then released by a wash with
9:1 (v/v) MeOH/water (60 L). Slowly adding the product
fractions (20.4 kg) to stirred ethyl acetate (240 kg) at 2−8 °C
resulted in the precipitation of 1. This technique provided 551
g of 1 (88.2% pure) as a white amorphous powder in a 35%
isolated yield. Both quantity and quality were adequate for the
exploratory toxicity studies.
Figure 3. Structure of side product 25.
reaction mixture was treated with a solution of trifluoroethanol
(45 equiv) in MTBE in an attempt to quench excess BCl3 and
break up 1−boron complex before filtration.14 The collected
solid was set in filter under nitrogen-blowing for a long time
and then triturated in MTBE to further remove acidic volatiles.
Thus, a 2.27 kg-scale reaction afforded 1.7 kg of crude 1 (63%
pure) with a 115% mass recovery for the first scale-up. Further
experiments turned out that the reverse addition of the 21
solution in dichloromethane to a precooled (0−10 °C) 1 M
BCl3 solution in n-heptane provided a slightly cleaner reaction
with improved product stability that eliminated the need of
trifluoroethanol treatment. Thus, the reaction slurry was
directly filtered, and the collected solid was blown with
nitrogen for 16 h producing crude 1 (∼70% pure) with a
∼150% mass recovery on a multiple 10 g scale.
Crystallization Development of 1 for Further Scale-Up.
With this “fit-for-purpose” procedure secured to prepare
Monobactam 1, our attention shifted to deliver 5 kg of 1 for
the regulatory toxicity studies and Phase 1 clinical trials.
Therefore, we needed to develop a practical purification
procedure that could provide 1 with increased purity, ideally as
a crystalline solid. To achieve this goal, we further purified a
small amount of the exploratory toxicity lot (88.2% pure of 1)
by reverse-phase chromatography, CG161M resin treatment,
and lyophilization. Pure amorphous 1 (∼2 g, 97.3% pure) was
obtained to study stability and solubility in water under
Purification of 1 for the First Delivery. The first delivery of
monobactam 1 targeted 500 g of the API with a purity
acceptance of >85% for exploratory toxicity studies. While this
E
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
different pH values. Surprisingly, 1 is quite stable in the acidic
aqueous solution from pH 1 to 4, and opening of the
monobactam ring to generate 25 was relatively slow (Table 1).
and elemental analysis reveal that Form 1 is Monobactam 1
pentahydrate (1·5H2O).
It had been observed that the Amberchrom CG161M resin
was able to purge impurities and purify 1 if the ratio of water
and methanol (or acetonitrile) was properly adjusted in the
development of the resin-capture−release technique to remove
water. Combining this observation and the crystallization
knowledge, we developed a practical isolation protocol that
utilizes CG161M resin to purify crude 1 and then crystallizes
the final product (1·5H2O) from a buffered aqueous solution
at pH ∼1.12 on gram scale. Thus, the feed slurry was prepared
by dissolving crude 1 (∼70% pure) in 1:2 water/acetonitrile,
adjusting pH to ∼3 with aqueous sodium formate, adding the
Amberchrom CG161M resin (100 wt %), and then removing
acetonitrile under vacuum. The purification was then achieved
by loading this feed onto an Amberchrom CG161M resin
column (800 wt %) and eluting with 15:1 v/v of aqueous
phosphate buffer (pH 3.1) and acetonitrile. The fractions
containing >85% pure 1 were combined, concentrated to
remove acetonitrile, and then loaded onto a second CG161M
resin column (400 wt %) for the resin-capture−release
manipulation. After water wash and nitrogen blow to remove
the bulk amount of water and phosphate salts from the
column, the captured product was eluted by 5:2 (v/v)
acetonitrile/water (∼40 volumes). The removal of acetonitrile
under vacuum generated a pH ∼3 aqueous solution of 1
(∼90% pure) in a ∼5 wt % concentration, which was set for
crystallization. Adding 1 N hydrochloric acid to adjust the
aqueous solution of 1 to pH 1.12, followed by stirring at 22 °C
for 24 h with intermittently adding either dilute aqueous HCl
or NaOH (as needed) to maintain the pH between 1.10 and
1.15, resulted in a white slurry. Filtration, followed by a water
wash and drying at 25 °C under vacuum to a constant weight,
produced crystalline 1·5H2O (Form 1, ∼94% pure).
Triturating this solid in diluted aqueous hydrochloric acid at
pH ∼1.12 upgraded the purity to 97.5% in the same form. The
yield of the final step (from the global deprotection reaction to
the crystallization) was 46%. Thus far, we enabled the synthesis
with this purification/crystallization procedure.
Table 1. Stability of 1 in Water vs pH (22 °C)
24 h purity of 1
(25)
48 h purity of 1
(25)
7 days purity of 1
(25)
pH
1
1.5
2
3
4
95.84% (1.51%)
95.68% (1.46%)
95.62% (1.41%)
95.71% (1.41%)
95.60% (1.71%)
95.85% (1.62%)
95.55% (1.55%)
95.38% (1.44%)
95.47% (1.39%)
95.39% (1.73%)
94.17% (2.56%)
93.89% (2.19%)
92.92% (1.92%)
92.86% (1.85%)
93.37% (2.09%)
T = 0 control 97.25% (1.19%)
On the other side, the aqueous solubility of 1 quickly
decreased from pH 4 to 3 and stayed flat at ∼2 mg/mL in
the range of pH 1−3 (Figure 4). This solubility pattern is likely
Figure 4. Solubility of 1 in water vs pH (22 °C).
attributed to the zwitterionic nature of the molecule, which
gave a calculated isoelectric point at pH 0.77.17 With this
knowledge, the crystallization of 1 was screened in buffered
aqueous solutions at pH 0.7, 0.8, 0.9, 1.1, 1.3, 1.5, 2.0, and 2.5
(pH adjusted by various ratios of 1 N hydrochloric acid and 1
N KH2PO4 solution) in a 5 wt % concentration at 22 °C. After
48 h, crystalline solids were observed in two vials at pH 1.1 and
1.3, and the isolated white solid 1 owned the same polymorph
(Form 1, see Figure 5) with ∼98% purity. KF measurement
CONCLUSIONS
■
A new synthesis of Monobactam 1 from three components (2,
3, and 4) was designed, and the corresponding process was
developed as follows: forming the unsymmetrical urea linker
Figure 5. PXRD of 1·5H2O (Form 1).
F
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
through CDI activation, removing the DMB group by oxidative
cleavage, deprotecting N-Boc group with TFA, constructing
the amide bond through two activation methods, introducing
the N-sulfonic acid, and finally removing all remaining
protecting groups in one step. In dealing with the
intermediates lacking crystallinity, isolations with partial
purification were frequently accomplished by precipitation en
route. The Amberchrom CG161M resin was successfully used
not only as a resin-capture−release technique to remove the
bulk amount of water but also as an efficient purification tool
to upgrade the API quality. More importantly, a process that
isolated the API crystal as a pentahydrate form (1·5H2O) from
the acidic aqueous solution was developed based on studies of
the stability and solubility of 1 at various pH values. The initial
process work facilitated the delivery of 500 g of amorphous 1
for the exploratory toxicity studies, and the identification of the
API crystallization process enabled the synthesis for further
scale-up with improved quality, stability, and throughput.
at 20 °C with ethyl acetate rinse (3 L). The white slurry was
filtered with an n-heptane wash (6 L). The solid was dried at
50 °C under vacuum with nitrogen sweep over 6 h to afford 7
(4.90 kg, 6.71 mol) as a white powder in a 94.1% yield. HPLC
purity: 95.9%. HRMS (ESI) m/z: calculated for C39H46N5O9
[M + H]+ 728.3296; observed 728.3288. 1H NMR (400 MHz,
DMSO-d6): δ 8.13 (s, 1H), 7.65 (d, J = 9.4 Hz, 1H), 7.54 (m,
2H), 7.38−7.47 (m, 8H), 7.10 (d, J = 8.2 Hz, 1H), 6.54 (m,
2H), 6.47 (dd, J = 8.2/2.3 Hz, 1H), 6.14 (m, 1H), 6.11 (s,
1H), 5.30 (s, 2H), 5.03 (s, 2H), 4.75 (dd, J = 9.4/4.3 Hz, 1H),
4.41 (d, J = 14.8 Hz, 1H), 4.22 (m, 1H), 4.02 (d, J = 14.8 Hz,
1H), 3.75 (s, 3H), 3.74 (s, 3H), 3.51 (m, 1H), 3.30 (m, 1H),
3.14 (m, 1H), 1.39 (s, 9H). 13C NMR (100 MHz, DMSO-d6):
δ 170.09, 166.96, 160.71, 158.47, 157.99, 155.61, 147.21,
146.91, 136.94, 133.36, 130.71, 130.06, 129.19, 128.91, 128.59,
123.77, 116.30, 111.17, 104.89, 98.84, 80.88, 79.13, 71.25,
58.71, 57.26, 55.91, 55.66, 37.69, 28.59.
Preparation of tert-Butyl ((2R,3S)-2-((3-((1,5-Bis-
(benzyloxy)-4-oxo-1,4-dihydropyridin-2-yl)methyl)ureido)-
methyl)-4-oxoazetidin-3-yl)carbamate (11). To a 200 L
reactor, equipped with a pH probe, a dip pipe, and a pH
titrator, were charged acetonitrile (70.0 L), 7 (4.67 kg, 6.41
mol), potassium persulfate (6.94 kg, 25.7 mol), K2HPO4 (5.03
kg, 28.9 mol), and water (37.3 L). After setting the nitrogen
purge rate at >2 standard cubic feet per minute, the mixture
was heated to 72 °C over 30 min. The mixture was stirred at
70−72 °C for ∼2 h for reaction completion. During the period,
part (as needed) of a stock solution of K2HPO4 (5.50 kg, 28.9
mol) in water (64.3 L) was added in a combination manner
from the stock tank (main) and the pH titrator (supplemental)
to maintain the pH of the reaction mixture in the range of 6.3−
7.0 when the pH was dropped to 6.3. The fact that the pH
stopped dropping indicated the reaction completion, which
was further confirmed by HPLC analysis. The mixture was
cooled to 20 °C over 30 min and filtered with acetonitrile
washes (2 × 10 L) to remove most inorganic precipitates. The
combined filtrates were concentrated under vacuum at ∼35 °C
to a residual volume of ∼83 L. Water (23.3 L) and ethyl
acetate (56.0 L) were added. The mixture was concentrated
again under vacuum at ∼40 °C to a residual volume of ∼98 L.
The mixture was then cooled to 20 °C, and ethyl acetate (93.4
L) was added. The resulting slurry was stirred for 20 min and
then filtered. The cake was washed with water (37.3 L) and
ethyl acetate (18.7 L) and dried at 60 °C under vacuum with
nitrogen sweep for 8 h to give crude 11 (3.08 kg). This crude
was charged back to the reactor, followed by the addition of
water (30.8 L) and ethyl acetate (30.8 L). The slurry was
stirred at 20 °C for 1 h and filtered with a wash of water (5 L)
and ethyl acetate (10 L). The solid was dried at 60 °C under
vacuum with nitrogen sweep for 12 h to afford 11 (2.24 kg,
3.88 mol) as an off-white powder in a 61% yield. HPLC purity:
96.6%. HRMS (ESI) m/z: calculated for C30H36N5O7 [M +
H]+ 578.2615; observed 578.2601. 1H NMR (400 MHz,
DMSO-d6): δ 8.21 (s, 1H), 8.00 (s, 1H), 7.63 (d, J = 9.8 Hz,
1H), 7.32−7.57 (m, 10H), 6.51 (m, 1H), 6.11 (m, 1H), 5.96
(s, 1H), 5.26 (s, 2H), 5.00 (s, 2H), 4.81 (dd, J = 9.4/4.4 Hz,
1H), 4.19 (m, 1H), 3.65 (m,1H), 3.16 (m, 2H), 1.40 (s, 9H).
13C NMR (100 MHz, DMSO-d6): δ 171.19, 168.28, 158.14,
155.63, 147.07, 146.83, 137.14, 133.48, 130.68, 129.99, 129.18,
128.89, 128.55, 128.51, 123.37, 111.27, 80.62, 79.06, 71.08,
59.52, 53.27, 41.43, 37.59, 28.61.
EXPERIMENTAL SECTION
■
General. All commercially available materials and solvents
were used as received, unless otherwise stated. All reactions
were executed under a nitrogen atmosphere. Reaction
temperatures were measured internally, unless indicated
otherwise. Achiral high-performance liquid chromatography
(HPLC) analyses were carried out using an XBridge C8
column (3.0 mm × 50 mm, 2.5 μm); column temperature 45
°C; flow rate 2.0 mL/min; detection UV 210 nm; mobile
phase: acetonitrile (solvent A), 0.05% methanesulfonic acid in
water with 10 mM sodium octylsulfonate (solvent B); linear
gradient elution (12 min): 0−8.3 min increasing solvent A
from 5% to 95, 8.3−10 min decreasing solvent A from 95 to
5%, and 10−12 min holding solvent A at 5%. Purity was
reported as the area percentage of the main peak.
Preparation of tert-Butyl ((2R,3S)-2-((3-((1,5-Bis-
(benzyloxy)-4-oxo-1,4-dihydropyridin-2-yl)methyl)ureido)-
methyl)-1-(2,4-dimethoxybenzyl)-4-oxoazetidin-3-yl)-
carbamate (7). To a 100 L reactor were charged K3PO4 (2.27
kg; 10.70 mol) and water (8.34 L). A suspension of 4 mesylate
salt (3.28 kg; 7.13 mol) in dichloromethane (20.8 L) was
added with a dichloromethane rinse (5.2 L) while maintaining
the temperature between 15 and 25 °C. The mixture was
stirred at 20 °C for 30 min. The bottom organic layer was
separated, and the top aqueous layer was back-extracted with
dichloromethane (13 L). The combined organic layers were
dried with MgSO4 (4.9 kg) and filtered with a dichloro-
methane rinse (3.2 L) to give a solution of 4 free base in
dichloromethane (KF < 0.2%). This solution was added to a
solution of CDI (1.70 kg, 10.50 mol) in dichloromethane (26
L) over 3 h with a dichloromethane rinse (5.0 L) while
maintaining the temperature between 15 and 25 °C. The
mixture was stirred at 20 °C for 1 h. The reaction completion
was monitored by proton NMR. 2 (2.61 kg, 7.13 mol) was
then added, followed by DMF (6.0 L). The mixture was stirred
at 20 °C for 4 h and then at 40 °C for 17 h. Upon reaction
completion, most of dichloromethane was distilled off under
vacuum to a residual volume of ∼13 L. Ethyl acetate (56 L)
was added. The mixture was washed with a 10 wt % aqueous
citric acid solution (58 kg) and 5 wt % aqueous NaCl solution
(2 × 58 kg). The organic layer was then dried with MgSO4
(4.9 kg), and the filtrates were concentrated under vacuum at
<35 °C to a residual volume of ∼28 L. The concentrated
solution was added to a stirred n-heptane (108 L) over 30 min
Preparation of 1-(((2R,3S)-3-Amino-4-oxoazetidin-2-yl)-
methyl)-3-((1,5-bis(benzyloxy)-4-oxo-1,4-dihydropyridin-2-
G
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
yl)methyl)urea Trifluoroacetic Acid Salt (14). To a 100 L
reactor were charged trifluoroacetic acid (36.81 kg, 322.9 mol)
and anisole (8.05 kg, 56.0 mol). The solution was cooled to −7
°C. 11 (3.73 kg; 6.46 mol) was added. The mixture was stirred
at −5 to −10 °C for 3 h. Upon reaction completion, the
mixture was slowly transferred to a cold (10 °C) mixture of
diatomaceous earth (8.21 kg) in MTBE (112.0 L) in a 200 L
reactor while maintaining the temperature in the receiving
reactor around 10 °C. The slurry was stirred at 10 °C for 30
min and filtered. The cake was washed with MTBE (69.4 L)
and dried with nitrogen-blowing for 1 h. This crude material
was charged back to the 200 L reactor and stirred with MTBE
(112 L) at 20 °C for 30 min and then filtered with an MTBE
wash (69.4 L). The solid was dried with nitrogen-blowing for
12 h to a constant weight to give a mixed white solid (11.32
kg), which contained 14 (3.11 kg, 5.26 mol) and diatomaceous
earth (8.21 kg) in an 81% yield. HPLC purity: 86.5%. MS
(ESI) m/z: calculated for C25H28N5O5 [M + H]+ 478.2090;
observed 478.2. 1H NMR (400 MHz, DMSO-d6): δ 8.81 (brs,
2H), 8.70 (s, 1H), 8.43 (s, 1H), 7.34−7.54 (m, 10H), 6.97 (t, J
= 6.0 Hz, 1H), 6.82 (t, J = 5.7 Hz, 1H), 6.41 (s, 1H), 5.36 (s,
2H), 5.09 (s, 2H), 4.47 (m, 1H), 4.28 (m, 2H), 3.73 (m, 1H),
3.36 (m, 1H), 3.22 (m, 1H).
Preparation of (Z)-2-(((1-(2-((tert-Butoxycarbonyl)-
amino)thiazol-4-yl)-2-((2,5-dioxopyrrolidin-1-yl)oxy)-2-
oxoethylidene)amino)oxy)-2-methylpropanoic Acid (18). To
a 20 L reactor were charged dichloromethane (11.5 L), 3 (1.15
kg, 2.68 mol), and N-hydroxysuccinimide (0.35 kg, 3.04 mol).
The solution was cooled to 2 °C. N,N′-Diisopropylcarbodii-
mide (0.37 kg, 2.93 mol) was added over 30 min while
maintaining the temperature between 0 and 10 °C. The
mixture was warmed to 22 °C and stirred for 3 h. Upon
reaction completion, the mixture was filtered through a 0.5 μm
polypropylene cartridge filter with a dichloromethane rinse
(2.0 L). The filtrates were concentrated under vacuum at <35
°C to a residual volume of ∼4 L. Methanol (12.0 L) was
added, and the mixture was concentrated under vacuum at <40
°C to a residual volume of ∼6.5 L. The mixture was then
cooled to 20 °C and stirred for 1 h. The slurry was filtered with
a methanol wash (2.30 L). The solid was dried at 40 °C under
vacuum with nitrogen sweep for 8 h to afford 18 (1.198 kg,
2.27 mol) as an off-white powder in an 85.0% yield. HPLC
purity: 98.1%. HRMS (ESI) m/z: calculated for C22H31N4O9S
[M + H]+ 527.1812; observed 527.1824. 1H NMR (400 MHz,
DMSO-d6): δ 11.97 (s, 1H), 7.57 (s, 1H), 2.90 (s, 4H), 1.49
(s, 6H), 1.48 (s, 9H), 1.37 (s, 9H). 13C NMR (100 MHz,
DMSO-d6): δ 171.79, 170.16, 162.10, 158.66, 153.51, 143.50,
139.98, 117.57, 83.77, 82.07, 81.23, 28.29, 28.05, 26.16, 23.89.
Preparation of 2-((((Z)-2-(((2R,3S)-2-((3-((1,5-Bis-
(benzyloxy)-4-oxo-1,4-dihydropyridin-2-yl)methyl)ureido)-
methyl)-4-oxoazetidin-3-yl)amino)-1-(2-((tert-
butoxycarbonyl)amino)thiazol-4-yl)-2-oxoethylidene)-
amino)oxy)-2-methylpropanoic Acid (19): Method A
through Succinimidyl Ester Activation. To a 75 L reactor
were charged acetonitrile (21.2 L) and the mixed solid (5.51
kg), which contained 14 (1.51 kg, 2.56 mol) and diatomaceous
earth (4.00 kg). Molecular sieves (3Å, 3.03 kg), 18 (1.38 kg,
2.62 mol), and DMAP (0.80 kg, 6.56 mol) were then added.
The mixture was heated to 40 °C and stirred for 16 h. Upon
reaction completion, the mixture was cooled to 20 °C. The
aqueous citric acid solution (2 wt %, 13 kg) was added to
adjust the pH to ∼6. The mixture was concentrated under
vacuum at ∼35 °C to a residual volume of ∼18 L. Ethyl acetate
(22.0 L) was added, and the mixture was filtered with a wash
of ethyl acetate (20 L) and water (7 L). The organic layer of
the combined filtrates was separated, washed with a 10 wt %
aqueous citric acid solution (17 kg) and 5 wt % aqueous NaCl
solution (2 × 10 kg), and then dried with MgSO4 (3.0 kg).
The mixture was filtered with an ethyl acetate wash (10 L).
The combined filtrates were concentrated under vacuum at
∼30 °C to a residual volume of ∼7 L. This solution was added
to n-heptane (36 L) over 30 min. The slurry was stirred at 20
°C for 1 h and filtered with a wash of 4:1 (v/v) heptane and
ethyl acetate (10 L). The solid was dried at 40 °C under
vacuum with nitrogen sweep for 8 h to afford crude 19 (1.92
kg, 81.0% HPLC purity) as a tan-colored powder in an 83%
yield.
Purification of 19 by Chromatography. A dynamic axial
compression column was packed with 85 kg of silica gel by
slurry in 2:1 (v/v) ethyl acetate-toluene and then equilibrated
with ethyl acetate (153 kg). The feed solution was prepared by
dissolving crude 19 (3.68 kg) in ethyl acetate (10.0 kg) in a
vessel at 20 °C. The entire feed was loaded onto the
equilibrated column with an ethyl acetate rinse (3.9 kg). The
column was eluted first with ethyl acetate (172 kg), followed
by 10:1 (w/w) ethyl acetate−isopropanol (355 kg) and then
5:3 (w/w) ethyl acetate−isopropanol (673 kg). The flow rate
was 3 L/min. The product was collected between 163 and 656
kg of the strongest mobile phase. An online UV detector set at
290 nm was used to monitor the elution and determine cut
points. All fractions were analyzed by offline HPLC. Product-
containing cuts were combined and concentrated via vacuum
distillation to a low stir volume. The partially concentrated
solution was transferred to a rotary evaporator and further
concentrated to dryness. The solids were redissolved in ethyl
acetate (9.0 kg). The concentrate was transferred to a pressure
canister and then added to n-heptane (62 kg) in a stirred vessel
over 50 min at 20−25 °C with an ethyl acetate rinse (4.0 kg)
to precipitate the product. The product slurry was stirred for 1
h at 20−25 °C and filtered on a pressure filter with a wash of n-
heptane (5.9 kg) and ethyl acetate (1.3 kg). The solid was
dried in the filter by nitrogen-blowing over 24 h to a constant
weight. Purified 19 (2.35 kg) was obtained as an off-white
powder in a 64% yield. HPLC purity: 97.8%. HRMS (ESI) m/
z: calculated for C43H53N8O11S [M + H]+ 889.3555; observed
889.3534. 1H NMR (400 MHz, DMSO-d6): δ 11.85 (brs, 1H),
9.27 (d, J = 8.6 Hz, 1H), 8.41 (s, 1H), 8.09 (s, 1H), 7.54 (m,
2H), 7.32−7.48 (m, 8H), 7.29 (s, 1H), 6.67 (brs, 1H), 6.32
(brs, 1H), 6.08 (s, 1H), 5.29 (s, 2H), 5.19 (m, 1H), 5.03 (s,
2H), 4.24 (m, 2H), 3.78 (m,1H), 3.38 (m,1H), 3.19 (m, 1H),
1.43 (s, 9H), 1.41 (s, 6H), 1.40 (s, 9H). 13C NMR (100 MHz,
DMSO-d6): δ 172.38, 170.34, 167.06, 162.66, 160.96, 158.39,
153.45, 149.53, 147.03, 146.94, 142.47, 136.98, 133.37, 130.71,
130.04, 129.18, 128.91, 128.57, 123.69, 113.86, 111.19, 82.84,
81.75, 80.92, 80.83, 71.21, 57.64, 53.47, 42.04, 37.68, 28.28,
28.10, 24.27, 24.18.
Preparation of 2-((((Z)-2-(((2R,3S)-2-((3-((1,5-Bis-
(benzyloxy)-4-oxo-1,4-dihydropyridin-2-yl)methyl)ureido)-
methyl)-4-oxoazetidin-3-yl)amino)-1-(2-((tert-
butoxycarbonyl)amino)thiazol-4-yl)-2-oxoethylidene)-
amino)oxy)-2-methylpropanoic Acid (19): Method B
through CDMT In Situ Activation. To a 1 L round-bottom
flask were charged acetonitrile (664 mL) and the mixed solid
(251.9 g), which contained 14 (∼69.2 g, 0.117 mol) and
diatomaceous earth (∼182.7 g). The mixture was stirred at 20
°C for 15 min and then filtered with acetonitrile rinses (2 ×
H
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
330 mL). To the combined filtrates collected in a 2 L round-
bottom flask were added DMF (330 mL) and N-
methylmorpholine (38.7 mL, 0.35 mol). The mixture was
concentrated using a rotary evaporator under vacuum at <45
°C to a residual volume of ∼450 mL. A basic aluminum oxide
column (207 g, 50 mm D × 120 mm H) was packed with a
DMF rinse (200 mL). The above crude 14 solution in DMF
was loaded on top and eluted with DMF (500 mL). The
filtrates collected were 14 free base solution in DMF.
(m, 1H), 1.41 (s, 9H), 1.38 (s, 6H), 1.36 (s, 9H). 13C NMR
(100 MHz, DMSO-d6): δ 172.39, 163.96, 162.83, 162.55,
160.69, 158.14, 153.39, 149.31, 149.10, 146.04, 142.47, 135.83,
132.60, 130.97, 130.42, 129.30, 129.08, 128.99, 128.80, 126.18,
113.56, 110.70, 82.90, 82.44, 81.81, 80.86, 72.20, 59.24, 56.03,
41.67, 38.29, 28.26, 28.09, 24.24, 24.21.
Preparation of 2-((((Z)-1-(2-Aminothiazol-4-yl)-2-
(((2R,3S)-2-((3-((1,5-dihydroxy-4-oxo-1,4-dihydropyridin-2-
yl)methyl)ureido)methyl)-4-oxo-1-sulfoazetidin-3-yl)-
amino)-2-oxoethylidene)amino)oxy)-2-methylpropanoic
Acid (Crude 1): Using 1 M BCl3 Solution in Dichloro-
methane. To a 200 L reactor were charged dichloromethane
(45.4 L) and 21 (2.27 kg, 2.34 mol). The solution was
concentrated under vacuum at ∼18 °C to a residual volume of
∼24 L to reach KF < 0.1%. The mixture was then cooled to −5
°C, and the reactor was connected to an aqueous NaOH
scrubber. The BCl3 solution (1 M) in dichloromethane (34.20
kg, 23.42 mol) was added over 60 min while maintaining the
temperature between −5 and 5 °C. The mixture was warmed
to 15 °C over 45 min and stirred for 30 min. Upon reaction
completion, the mixture was cooled to −15 °C. A precooled
(−15 °C) solution of trifluoroethanol (10.55 kg, 105.41 mol)
in MTBE (45.4 L) was added over 30 min while maintaining
the temperature below −10 °C. The slurry was warmed to 0
°C over 15 min, stirred for 30 min, and then filtered with an
MTBE wash (45.4 L). The cake was dried with nitrogen-
blowing for 3 h and then charged back to the reactor. MTBE
(45.4 L) was added. The slurry was stirred at 20 °C for 1 h and
filtered with an MTBE wash (20 L). The cake was dried with
nitrogen-blowing for 12 h to afford crude 1 (1.72 kg, 63%
HPLC purity) as an off-white powder with a 115% mass
recovery.
To a 3 L three-neck round-bottom flask were charged 3
(42.18 g, 98.22 mmol), CDMT (17.24 g, 98.22 mmol), and
ethyl acetate (664 mL). N-Methylmorpholine (23.2 mL, 210.5
mmol) was added. The mixture was stirred at 20 °C for 60
min. The reaction completion was monitored by HPLC with
benzylamine derivatization. The above 14 free base solution in
DMF was added over 15 min via addition funnel. The mixture
was stirred at 20 °C for 16 h. Upon reaction completion, ethyl
acetate (1 L) was added. The mixture was washed with a 5 wt
% aqueous NaCl solution (830 mL), and the aqueous layer was
back-extracted with ethyl acetate (1 L). The combined organic
layers were then washed with a 5 wt % aqueous NaCl solution
(2 × 830 mL), 10 wt % aqueous citric acid solution (830 mL),
10 wt % aqueous K3PO4 solution (830 mL), and saturated
aqueous NaCl solution (830 mL). The organic layer was then
concentrated using a rotary evaporator at ∼50 °C to a residual
weight of ∼300 g, which was filtered with ethyl acetate rinses
(2 × 25 mL). The combined filtrates were added slowly to n-
heptane (1.5 L) over 30 min. The slurry was stirred at 20 °C
for 1 h and filtered with a wash of 5:1 (v/v) n-heptane and
ethyl acetate (400 mL). The solid was dried under vacuum at
50 °C for 15 h to afford 19 (72.36 g, 81.39 mmol) as a light
tan powder in a 69% yield. HPLC purity: 96.1%.
Preparation of (2R,3S)-2-((3-((1,5-Bis(benzyloxy)-4-oxo-
1,4-dihydropyridin-2-yl)methyl)ureido)methyl)-3-((Z)-2-(((1-
(tert-butoxy)-2-methyl-1-oxopropan-2-yl)oxy)imino)-2-(2-
((tert-butoxycarbonyl)amino)thiazol-4-yl)acetamido)-4-ox-
oazetidine-1-sulfonic Acid (21). To a 100 L reactor were
charged DMF (23.0 L) and 19 (2.30 kg, 2.59 mol). The
solution was cooled to 15 °C, and sulfur trioxide−DMF
complex (3.96 kg, 25.87 mol) was added. The mixture was
stirred at 20 °C for 30 min. Upon reaction completion, the
mixture was cooled to 0 °C and transferred with a DMF rinse
(2.3 L) to a precooled (2 °C) mixture of ethyl acetate (69 L)
and 5 wt % aqueous NaCl solution (46 kg) in a 200 L reactor
over 30 min. The organic layer was separated, and the aqueous
layer was back-extracted with ethyl acetate (23 L). The
combined organic layers were washed with a 5 wt % aqueous
NaCl solution (3 × 46 kg) and then saturated aqueous NaCl
solution (26 kg). MgSO4 (6.90 kg) was added to the organic
layer, and the mixture was stirred for 1 h (for both drying and
converting 22 to 21). Upon reaction completion, the mixture
was filtered with an ethyl acetate rinse (15 L). The combined
filtrates were concentrated under vacuum at ∼30 °C to a
residual volume of ∼8 L. MTBE (69 L) was added to over 1 h.
The slurry was stirred at 20 °C for 1 h and filtered with an
MTBE wash (9.2 L). Nitrogen-blowing the cake for 12 h
afforded 21 (2.19 kg, 2.26 mol) as an off-white powder in an
87% yield. HPLC purity: 89.8%. MS (ESI) m/z: calculated for
C43H53N8O14S2 [M + H]+ 969.3123; observed 969.3. 1H NMR
(400 MHz, DMSO-d6) δ 9.27 (d, J = 8.8 Hz, 1H), 8.86 (s,
1H), 7.57 (dd, J = 5.7/2.0 Hz, 2H), 7.36−7.50 (m, 10H), 7.26
(s, 1H), 6.91 (s, 1H), 6.38 (m, 1H), 5.49 (s, 2H), 5.20 (s, 2H),
5.19 (m, 1H), 4.38 (m, 2H), 3.93 (m,1H), 3.82 (m,1H), 3.10
Purification of 1 by Reversed-Phase Chromatogra-
phy. A dynamic axial compression column was packed with 9
kg of Kromasil C-18 by slurry in isopropanol, rinsed with
purified water (low iron content, 40 L), and equilibrated with
mobile phase (acetonitrile-purified water and 0.2% formic acid
modifier, 145 kg). The column feed solution was prepared by
dissolving crude 1 (1.70 kg) in 1 M aqueous ammonium
bicarbonate (10.4 L) and DMF (2.2 L) in a stirred vessel
placed in an ice bath, and the solution was filtered. For each
run, the feed solution (148 mL) was loaded onto the column,
and the column was eluted with the mobile phase (acetonitrile-
purified water and 0.2% formic acid modifier) using a reverse-
phase gradient algorithm. The flow rate was 4.2 L/min. An
online UV detector set at 320 nm was used to monitor the
elution and determine cut points. All fractions were analyzed
by offline HPLC. Product-containing cuts were combined in a
stirred vessel and concentrated via vacuum distillation until
acetonitrile was removed.
A dynamic axial compression column was packed with 22 L
of the Amberchrom CG161M resin by slurry in 4:1 (v/v)
ethanol−water and rinsed with methanol (60 L) and then
purified water (60 L). The product-containing partial
concentrate was loaded onto the column. The column was
washed with purified water, purged with nitrogen to remove
free water, and eluted with 9:1 (v/v) methanol−purified water
(60 L). Product-containing cuts (20.4 kg) were combined in a
pressure canister and then added to ethyl acetate (240 kg) with
a methanol rinse (4.0 kg) in a stirred vessel over 60 min at 2−8
°C to precipitate the product. Additional ethyl acetate (63 kg)
was added, and the slurry was stirred for 40 min at 2−8 °C and
filtered. The filter cake was washed with ethyl acetate (2 × 9.0
I
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
kg) and blown with nitrogen for 8 h. The solid was dried at 30
°C under vacuum with nitrogen sweep for 8 h to afford 1 (551
g, 0.814 mol, 88.2% HPLC purity) as a white powder in a
34.8% yield.
DMSO-d6) δ 9.22 (d, J = 8.8 Hz, 1H), 8.18 (s, 1H), 7.37 (brs,
1H), 7.22 (brs, 1H), 7.00 (s, 1H), 6.74 (s, 1H), 6.34 (m, 1H),
5.17 (dd, J = 8.7/5.8 Hz, 1H), 4.33 (m, 2H), 3.95 (m, 1H),
3.63 (m, 1H), 3.22 (m, 1H), 1.39 (s, 3H), 1.37 (s, 3H). 13C
NMR (100 MHz, DMSO-d6): δ 175.18, 169.32, 162.85,
162.59, 158.24, 157.54, 148.93, 145.74, 144.57, 141.51, 127.90,
110.14, 109.76, 82.39, 58.99, 56.15, 41.19, 38.64, 24.38, 24.25.
KF: calculated for 1·5H2O, 12.4 wt % of water; found 12.1 wt
%. Elemental analysis: calculated for 1·5H2O, C: 33.24%, H:
4.74%, N: 15.51%; found C: 32.86%, H: 4.69%, N: 15.38%.
PXRD: Form 1.
Preparation of 2-((((Z)-1-(2-Aminothiazol-4-yl)-2-
(((2R,3S)-2-((3-((1,5-dihydroxy-4-oxo-1,4-dihydropyridin-2-
yl)methyl)ureido)methyl)-4-oxo-1-sulfoazetidin-3-yl)-
amino)-2-oxoethylidene)amino)oxy)-2-methylpropanoic
Acid Crude (1): Using 1 M BCl3 Solution in n-Heptane. To a
250 mL three-necked round-bottom flask was charged 1 M
BCl3 solution in n-heptane (74.3 mL, 74.3 mmol). The
solution was cooled to 10 °C. A solution of 21 (9.00 g, 9.30
mmol) in dichloromethane (90 mL) was added via addition
funnel over 15 min while maintaining the temperature between
8 and 12 °C. The mixture was stirred at 10 °C for 30 min.
Upon reaction completion, the slurry was filtered under
nitrogen pressure and washed by 1:1 dichloromethane and n-
heptane three times (3 × 90 mL). The cake was dried with
nitrogen-blowing for 16 h to afford 70% pure crude 1 (8.84 g)
as a yellow powder with a 150% mass recovery. The material
was stored in a refrigerator for use.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: y_tao@yahoo.com (Y.T.).
*E-mail: manjinder.lall@pfizer.com (M.S.L.).
■
ORCID
Yong Tao: 0000-0001-9773-7807
Manjinder S. Lall: 0000-0002-2882-8075
Notes
The authors declare no competing financial interest.
Purification of 1 through Amberchrom CG161M
Resin Column and Crystallization of 1·5H2O. To a 250
mL round-bottom flask were charged crude 1 (5.60 g, 70%
purity, estimated ∼4 mmol) and acetonitrile (60 mL). The
slurry was cooled to 10 °C with ice/water bath. Water (30 mL)
was added slowly, followed by sodium formate (3.40 g, 50
mmol) portion-wise to adjust the pH to 2.9−3.1 while
controlling the temperature below 20 °C. The Amberchrom
CG161M resin (7.2 g, prewashed with water) was added.
Acetonitrile was removed by a rotary evaporator with bath
temperature <30 °C. The brown slurry was loaded onto an
Amberchrom CG161M resin column (60 g, 32 mm D × 110
mm H) with a water rinse (200 mL). The column was then
eluted with 2.8 L of a 15:1 (v/v) aqueous phosphate buffer
(pH 3.1, prepared by dissolving 1.0 g of KH2PO4 and 60 mg of
85% phosphoric acid to 1 L water) and acetonitrile. The
fractions that contained >85% pure 1 (∼1.1 L) were
combined. Acetonitrile was removed by a rotary evaporator
with bath temperature <30 °C, and the aqueous solution
(∼91% purity) was loaded onto an Amberchrom CG161M
resin column (30 g, 32 mm D × 55 mm H) and washed with
water (150 mL). The column was blown with nitrogen for 10
min and then eluted with 5:2 (v/v) acetonitrile and water (330
mL). The filtrates were concentrated to a residual volume of
∼70 mL by a rotary evaporator with bath temperature <35 °C.
Hydrochloric acid (1 N, 5 mL) was added slowly to adjust the
pH to a constant 1.12. The seed was added, and the mixture
was stirred at 21 °C for 24 h while an additional amount (∼2
mL in total) of 1 N hydrochloric acid was added intermittently
to maintain pH ∼1.12. The resulting slurry was filtered with
water rinses (2 × 4 mL). After drying at 25 °C for 48 h, 94.5%
pure 1·5H2O (2.04 g) was obtained as an off-white powder.
This 1·5H2O and water (30 mL) were added to a 100 mL
round-bottom flask, and the pH of the slurry was measured at
2.23. Hydrochloric acid (1 N, 3.2 mL) was added slowly to
adjust the pH to a constant 1.12. The slurry was stirred at 21
°C for 24 h and filtered with water rinses (2 × 4 mL). After
drying at 25 °C for 48 h, 1·5H2O (Form 1, 1.96 g, 2.71 mmol)
was obtained as an off-white powder in a 45.8% yield (from
reaction to trituration). HPLC purity: 97.45% (with 0.62% of
25). HRMS (ESI) m/z: calculated for C20H25N8O12S2 [M +
H]+ 633.1033; observed 633.1045. 1H NMR (400 MHz,
ACKNOWLEDGMENTS
■
The authors acknowledge the extended Monobactam Team for
collaborative effort and helpful discussions. For their executive
and technical expertise, the authors extend their gratitude to
John Brennan, Hayden Thomas, Mark Mitton-Fry, Matt
Brown, Mark C. Noe, Jeffrey W. Raggon, and Joel T. Arcari.
The authors thank Dr. Bryan Li for manuscript review and
providing useful feedback. The staff members in Pfizer Groton
Kilo Lab and Separation Sciences Group are gratefully
recognized for their successful execution of this process.
REFERENCES
■
(1) (a) Brown, M. F.; Mitton-Fry, M. J.; Arcari, J. T.; Barham, R.;
Casavant, J.; Gerstenberger, B. S.; Han, S.; Hardink, J. R.; Harris, T.
M.; Hoang, T.; Huband, M. D.; Lall, M. S.; Lemmon, M. M.; Li, C.;
Lin, J.; McCurdy, S. P.; McElroy, E.; McPherson, C.; Marr, E. S.;
Mueller, J. P.; Mullins, L.; Nikitenko, A. A.; Noe, M. C.; O’Donnell, J.
P.; Penzien, J.; Plummer, M. S.; Schuff, B. P.; Shanmugasundaram, V.;
Starr, J. T.; Sun, J.; Tomaras, A.; Young, J. A.; Zaniewski, R. P.
Pyridone-Conjugated Monobactam Antibiotics with Gram-Negative
Activity. J. Med. Chem. 2013, 56, 5541−5552. (b) Brown, M. F.;
Mitton-Fry, M. J.; Han, S.; Lall, M.; Plummer, M. S.; Risley, H. L.;
Shanmugasundaram, V.; Starr, J. T. Preparation of Monobactams for
the Treatment of Bacterial Infections. WO2012073138
(2) Lall, M. S.; Tao, Y.; Arcari, J. T.; Boyles, D. C.; Brown, M. F.;
Damon, D. B.; Lilley, S. C.; Mitton-Fry, M. J.; Starr, J.; Stewart, A. M.;
Sun, J. Process Development for the Synthesis of Monocyclic β-
Lactam Core 17. Org. Process Res. Dev. 2018, 22, 212−218.
(3) (a) Handa, V. K.; Gupta, S. M.; Maheshwari, N.; Rohatgi, A.
Process for the Preparation of Crystalline (Z)-2-(2-Tert-butoxycarbo-
nylprop-2-oxyimino)-2-(2-triphenylmethylaminothiazol-4-yl)acetic
Acid in Association with N,N-Dimethylformamide. U.S. Patent
US6214997 (b) Liao, X.; Pearson, N. D.; Pendrak, I.; Sano, M.
Preparation of Cephalosporins as Antibacterial Compounds.
WO2013052568
(4) (a) Di Fabio, R.; Griffante, C.; Alvaro, G.; Pentassuglia, G.; Pizzi,
D. A.; Donati, D.; Rossi, T.; Guercio, G.; Mattioli, M.; Cimarosti, Z.;
Marchioro, C.; Provera, S.; Zonzini, L.; Montanari, D.; Melotto, S.;
Gerrard, P. A.; Trist, D. G.; Ratti, E.; Corsi, M. Discovery Process and
Pharmacological Characterization of 2-(S)-(4-Fluoro-2-
methylphenyl)piperazine-1-carboxylic Acid [1-(R)-(3,5-Bis-
trifluoromethylphenyl)ethyl]methylamide (Vestipitant) as a Potent,
Selective, and Orally Active NK1 Receptor Antagonist. J. Med. Chem.
J
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
2009, 52, 3238−3247. using triphosgene: (b) Lukin, K.; Hsu, M. C.;
Chambournier, G.; Kotecki, B.; Venkatramani, C. J.; Leanna, M. R.
Development of a Large Scale Asymmetric Synthesis of Vanilloid
Receptor (TRPV1) Antagonist ABT-102. Org. Process Res. Dev. 2007,
11, 578−584 through succinimidyl carbamate: . (c) Ion, A.;
Parvulescu, V.; Jacobs, P.; De Vos, D. Synthesis of Symmetrical or
Asymmetrical Urea Compounds from CO2 via Base Catalysis. Green
Chem. 2007, 9, 158−161 Using CO2: . (d) Zheng, N.; Armstrong, J.
D.; Eng, K. K.; Keller, J.; Liu, T.; Purick, R.; Lynch, J.; Hartner, F. W.;
Volante, R. P. A Convergent Asymmetric Synthesis of a Growth
Hormone Secretagogue. Tetrahedron: Asymmetry 2003, 14, 3435−
3446 through isocyanate: . (e) Enders, D.; Wortmann, L.; Ducker, B.;
Raabe, G. Asymmetric Synthesis of 1,2,3,4,5,6-Hexahydro-5-hydrox-
ypyrimidin-2-ones as Potential HIV-Protease Inhibitors. Helv. Chim.
Acta 1999, 82, 1195−1201 using bis(4-nitrophenyl) carbonate: .
(5) For some examples using CDI for activation, see: (a) Enomoto,
H.; Sawa, A.; Suhara, H.; Yamamoto, N.; Inoue, H.; Setoguchi, C.;
Tsuji, F.; Okamoto, M.; Sasabuchi, Y.; Horiuchi, M.; Ban, M.
Synthesis and Pharmacological Evaluation of 1,1,3-Substituted Urea
Derivatives as Potent TNF-α Production Inhibitors. Bioorg. Med.
Chem. Lett. 2010, 20, 4479−4482. (b) Maresca, K. P.; Hillier, S. M.;
Femia, F. J.; Keith, D.; Barone, C.; Joyal, J. L.; Zimmerman, C. N.;
Kozikowski, A. P.; Barrett, J. A.; Eckelman, W. C.; Babich, J. W. A
Series of Halogenated Heterodimeric Inhibitors of Prostate Specific
Membrane Antigen (PSMA) as Radiolabeled Probes for Targeting
Prostate Cancer. J. Med. Chem. 2009, 52, 347−357. (c) Waibel, M.;
Hasserodt, J. Diversity-Oriented Synthesis of a Drug-Like System
Displaying the Distinctive N→CO Interaction. J. Org. Chem. 2008,
73, 6119−6126. (d) Dauvergne, J.; Polidori, A.; Venien-Bryan, C.;
Pucci, B. Synthesis of a Hemifluorinated Amphiphile Designed for
Self-assembly and Two-dimensional Crystallization of Membrane
Protein. Tetrahedron Lett. 2008, 49, 2247−2250. (e) Elliott, G. I.;
Fuchs, J. R.; Blagg, B. S. J.; Ishikawa, H.; Tao, H.; Yuan, Z.-Q.; Boger,
D. L. Intramolecular Diels-Alder/1,3-Dipolar Cycloaddition Cascade
of 1,3,4-Oxadiazoles. J. Am. Chem. Soc. 2006, 128, 10589−10595.
(6) For some examples using CAN for DMB oxidative cleavage, see:
(a) Overman, L. E.; Osawa, T. A Convenient Synthesis of 4-
Unsubstituted β-Lactams. J. Am. Chem. Soc. 1985, 107, 1698−1701.
(b) Khan, M. A. Preparation of Spiro-lactam NMDA Receptor
Modulators for the Treatment of Depression and Related Disorders.
WO2018026763 (c) Shiozaki, M.; Ishida, N.; Sato, S. Synthesis of 2-
Amino-2-deoxy-α-D-altrofuranoside Derivatives from 2,3-O-Isopro-
pylidene-D-glyceraldehyde via Bicyclic β-Lactam Intermediates. Bull.
Chem. Soc. Jpn 1989, 62, 3950−3958. (d) Sugita, K.; Otsuka, M.; Oki,
H.; Haginoya, N.; Ichikawa, M.; Ito, M. Preparation of Tricyclic
Compounds Such as Pyrrolobenzoxazepine Derivatives and Analogs
Thereof for Treatment of Hypercholesteremia, Hyperlipemia, and
Arteriosclerosis. JP2008291018
(10) (a) Rode, A. B.; Son, S. J.; Hong, I. S. An Efficient One-pot N-
Acylation of Deoxy- and Ribo-cytidine Using Carboxylic Acids
Activated in situ with 2-Chloro-4,6-dimethoxy-1,3,5-triazine. Bull.
Korean Chem. Soc. 2010, 31, 2061−2064. (b) Coffey, D. S. 6-Chloro-
2,4-dimethoxy-sec-triazine. e-EROS Encycl. Reagents Org. Synth. 2001,
́
1−5. (c) Kaminski, Z. J.; Kolesinska, B.; Kolesinska, J.; Sabatino, G.;
Chelli, M.; Rovero, P.; Blaszczyk, M.; Glowka, M. L.; Papini, A. M. N-
Triazinylammonium Tetrafluoroborates. A New Generation of
Efficient Coupling Reagents Useful for Peptide Synthesis. J. Am.
Chem. Soc. 2005, 127, 16912−16920.
́
(11) (a) Fernandez-Resa, P.; Herranz, R.; Conde, S.; Arribas, E.
Stereoselective Synthesis of Cis-4-(substituted) Monobactams from
Ethyl Acetoacetate. J. Chem. Soc., Perkin Trans. 1 1989, 67−71.
(b) Kostova, M. B.; Myers, C. J.; Beck, T. N.; Plotkin, B. J.; Green, J.
M.; Boshoff, H. I. M.; Barry, C. E., III; Deschamps, J. R.; Konaklieva,
M. I. C4-Alkylthiols with Activity against Moraxella Catarrhalis and
Mycobacterium Tuberculosis. Bioorg. Med. Chem. 2011, 19, 6842−
6852. (c) Mewshaw, R. E.; Commons, T. J. Synthesis and
Antibacterial Activity of C-4 Substituted Monobactams. J. Antibiot.
1987, 40, 1563−1571.
(12) The role of MgSO4 was not clear. However, without MgSO4,
the hydrolysis of the 2nd N-sulfonic acid was very sluggish and
generated a lot of unidentified impurities.
(13) (a) Gaeta, A.; Molina-Holgado, F.; Kong, X. L.; Salvage, S.;
Fakih, S.; Francis, P. T.; Williams, R. J.; Hider, R. C. Synthesis,
Physical-chemical Characterization and Biological Evaluation of
Novel 2-Amido-3-hydroxypyridin-4(1H)-ones: Iron Chelators with
the Potential for Treating Alzheimer’s Disease. Bioorg. Med. Chem.
2011, 19, 1285−1297. (b) Piyamongkol, S.; Zhou, T.; Liu, Z. D.;
Khodr, H. H.; Hider, R. C. Design and Characterization of Novel
Hexadentate 3-Hydroxypyridin-4-one Ligands. Tetrahedron Lett.
2005, 46, 1333−1336.
(14) Gerrard, W.; Lappert, M. F. Interaction of Alcohols and Boron
Trichloride. J. Chem. Soc. 1951, 2545−2550.
(15) Amberchrom CG161M resin is a product of Dow Chemicals.
https://www.dow.com/en-us/markets-and-solutions/products/
A M B E R C H R O M C h r o m a t o g r a p h y R e s i n s /
AMBERCHROMCG161M (Accessed on Oct. 2, 2018).
(16) For references on resin to be used for “resin-capture-release”,
see: (a) Masque, N.; Galia, M.; Marce, R. M.; Borrull, F.
Functionalized Polymeric Sorbents for Solid-phase Extraction of
Polar Pollutants. J. High Resolut. Chromatogr. 1999, 22, 547−552.
(b) Boughtflower, B.; Lane, S.; Mutton, I.; Stasica, P. Generic
Compound Isolation Using Solid-phase Trapping as Part of the
Chromatographic Purification Process. Part 1. Proof of Generic
Trapping Concept. J. Comb. Chem. 2006, 8, 441−454.
(17) The isoelectric point of 1 was calculated based on MoKa pKa
prediction tool.
(7) For some examples using potassium persulfate for DMB
oxidative cleavage, see: (a) Mastalerz, H.; Menard, M.; Vinet, V.;
Desiderio, J.; Fung-Tomc, J.; Kessler, R.; Tsai, Y. An Examination of
O-2-Isocephems as Orally Absorbable Antibiotics. J. Med. Chem.
1988, 31, 1190−1196. (b) Nunez-Villanueva, D.; Bonache, M. A.;
Infantes, L.; Garcia-Lopez, M. T.; Martin-Martinez, M.; Gonzalez-
Muniz, R. Quaternary α,α-2-Oxoazepane α-Amino Acids: Synthesis
from Ornithine-Derived β-Lactams and Incorporation into Model
Dipeptides. J. Org. Chem. 2011, 76, 6592−6603.
(8) Ledwith, A.; Russell, P. J. Aromatic Chlorination by
Peroxodisulfate and Chloride Ions. Cation Radical Trapping by
Copper(II) Chloride. J. Chem. Soc., Perkin Trans. 2 1975, 1503−1508.
(9) (a) Colombo, R.; Mingozzi, M.; Belvisi, L.; Arosio, D.; Piarulli,
U.; Carenini, N.; Perego, P.; Zaffaroni, N.; De Cesare, M.; Castiglioni,
V.; Scanziani, E.; Gennari, C. Synthesis and Biological Evaluation (in
Vitro and in Vivo) of Cyclic Arginine-Glycine-Aspartate (RGD)
Peptidomimetic-Paclitaxel Conjugates Targeting Integrin αVβ3. J.
Med. Chem. 2012, 55, 10460−10474. (b) Westermann, J.; Schneider,
M.; Platzek, J.; Petrov, O. Practical Synthesis of a Heterocyclic
Immunosuppressive Vitamin D Analogue. Org. Process Res. Dev. 2007,
11, 200−205.
K
DOI: 10.1021/acs.oprd.9b00374
Org. Process Res. Dev. XXXX, XXX, XXX−XXX