Process research for the synthesis of RWJ-51204, a novel anxiolytic agent

Article abstract of DOI:10.1021/op990182l

RWJ-51204, the lead compound in our pyrido [1,2-α] benzimidazole (PBI) series, was shown to exhibit anxiolytic efficacy in animal models at doses which did not cause central nervous system side effects commonly observed with other anxiolytic agents. To prepare supplies of drug substance for early toxicological and clinical studies, we needed to develop a safe and scaleable synthesis. Our main focus was to improve the last two steps of the process which involved formation of the penultimate carboxamide intermediate followed by alkylation using potentially toxic chloromethyl ethyl ether. Due to safety issues concerning storage and handling of this reagent during the large scale synthesis, we investigated alternate routes to minimize potential exposure risks. The process research carried out for the final steps that led to the safe and cost-effective multi-kilogram synthesis of RWJ-51204 is described herein.

Full text of DOI:10.1021/op990182l

Organic Process Research & Development 1999, 3, 260−265
Process Research for the Synthesis of RWJ-51204, A Novel Anxiolytic Agent
Judith H. Cohen,* Cynthia A. Maryanoff, Stephen M. Stefanick, Kirk L. Sorgi, and Frank J. Villani, Jr.
The R.W. Johnson Pharmaceutical Research Institute, New Product Research, Chemical DeVelopment Department,
Spring House, PennsylVania 19477
Abstract:
is orally active, (2) it possesses a high level of anxiolytic
efficacy in animal models, (3) it shows very good in vitro
activity (IC50 ) 0.3 nM), and most importantly, (4) the large
therapeutic index in animal models and overall profile
suggests RWJ-51204 (1) might have efficacy in human
anxiety-related disorders at doses that should not cause CNS
side effects. For complete biological evaluation, we needed
to prepare multi-kilogram quantities of drug substance for
early development studies. This paper describes some of the
process research which led to the large scale, safe, and cost-
effective synthesis of highly pure RWJ-51204 (1).
RWJ-51204, the lead compound in our pyrido [1,2-a] benzimi-
dazole (PBI) series, was shown to exhibit anxiolytic efficacy in
animal models at doses which did not cause central nervous
system side effects commonly observed with other anxiolytic
agents. To prepare supplies of drug substance for early
toxicological and clinical studies, we needed to develop a safe
and scaleable synthesis. Our main focus was to improve the
last two steps of the process which involved formation of the
penultimate carboxamide intermediate followed by alkylation
using potentially toxic chloromethyl ethyl ether. Due to safety
issues concerning storage and handling of this reagent during
the large scale synthesis, we investigated alternate routes to
minimize potential exposure risks. The process research carried
out for the final steps that led to the safe and cost-effective
multi-kilogram synthesis of RWJ-51204 is described herein.
Introduction
Benzodiazepines, such as diazepam and clonazepam, are
recognized as very effective drugs for treating anxiety related
disorders.1 However, all of these currently marketed drugs
cause unwanted central nervous system (CNS) side effects
including motor and memory impairment, sedation, muscle
relaxation, and physical dependency. Recently, our Drug
Discovery group focused on identifying new therapeutic
agents that were as efficacious as the marketed benzodiaz-
epine drugs without the unwanted CNS side effects.
Initially, we evaluated the medicinal chemistry based
synthesis as outlined in Scheme 1. Treatment of com-
mercially available 4-fluoro-2-nitroaniline (2) with acryloni-
trile3 in the presence of an excess of Triton B afforded the
cyanoethyl derivative 3 in 93% yield. Selective reduction
of the nitro group via catalytic hydrogenation and subsequent
reaction with ethyl 3-amino-3-ethoxyacrylate hydrochloride
(5) led to the formation of 1-(2-cyanoethyl)-2-carbethoxy-
methyl-5-fluorobenzimidazole (6) in 75% yield. Ethanolysis
of the cyano group in ethanolic HCl gave the diester
intermediate 7 which upon treatment with freshly prepared
sodium ethoxide in ethanol underwent Dieckman cyclization
to afford the pyrido[1,2-a]benzimidazole ester 8 in 66% yield
for the combined two steps. Reaction with 2-fluoroaniline
led to the isolation of penultimate amide intermediate 9 in
58% yield after chromatographic purification. Deprotonation
of 9 with NaH in DMF followed by treatment with
chloromethyl ethyl ether (CMEE) in the presence of 15-
crown-5 afforded crude RWJ-51204 (1). The pure product
was obtained after column chromatography on silica gel in
35% yield from intermediate 8. The overall yield for the
eight-step synthetic route is 15%.
As a result of this effort, we identified a novel class of
compounds collectively referred to as pyrido[1,2-a]benzimi-
dazoles (PBIs) which were shown to be active in animal
models as anxiolytic agents.2 In vitro and in vivo data
indicate these compounds fit into the class of GABAA
receptor modulators commonly termed as partial agonists.
We selected the lead compound in this series, RWJ-51204
(1), as a development candidate for several reasons: (1) it
We determined that this synthetic sequence could be used
to prepare the desired drug substance with some modification.
We identified the following safety and chemical issues as
limitations to the large-scale synthesis of this product:
(1) Dunn, R. W.; Flanagan, D. M.; Martin, L. L.; Kerman, L. L.; Woods, A.
T.; Camacho, F.; Wilmot, C. A.; Cornfeldt, M. L.; Effland, R. C.; Wood,
P. L.; Corbett, R. Eur. J. Pharmacol. 1992, 214, 207.
(2) Reitz, A. B.; Jordan, A. D.; Sanfilippo, P. J.; Scott, M. K.; Vavouyios-
Smith, A. U.S. Patent 5,817,668, 1998.
(3) Acrylonitrile is a genotoxic material and is systemically absorbed through
the skin; therefore, proper personal protective equipment and monitoring
were used.
260
•
Vol. 3, No. 4, 1999 / Organic Process Research & Development
10.1021/op990182l CCC: $18.00 © 1999 American Chemical Society and The Royal Society of Chemistry
Published on Web 06/18/1999

Scheme 1
1. Reaction of ester intermediate 8 with 2-fluoroaniline
ethoxymethylating reagent due to the potential exposure risks
during storage and handling.
was carried out in xylene at elevated temperature. Due to
the poor solubilities of both 8 and 9, the reaction mixture
was heterogeneous, leading to difficulties with stirring,
reaction monitoring, and isolation of purified product. All
of these factors contributed to the low yield and throughput
for this step.
• In addition to the hazardous reagents, the conversion
of penultimate intermediate 9 to RWJ-51204 (1) was
inefficient, requiring the crude product to be purified by silica
gel chromatography. For large-scale synthesis, purification
by chromatography is neither practical nor cost-effective.
Our goal in devising an alternate route that avoided the
problems highlighted above was to improve the last two steps
of the synthesis starting with the ester intermediate 8.
2. Conversion of carboxamide 9 to the final product, RWJ-
51204 (1) was problematic for several reasons:
• The anion of intermediate 9 was generated with NaH
in DMF. It has been reported in the literature that combina-
tions of 15% NaH/DMF w/w pose a potential thermal
hazard.4,5 There have even been reports of plant-scale
incidents using these reagents where they observed the onset
of an exotherm at 40 °C followed by rapid self-heating and
eventually explosion.5
• The ethoxymethylation was carried out in the presence
of 15-crown-5. This reagent is very toxic as well as an
irritant. For the final large-scale process, we needed to
eliminate this hazardous reagent.
• The alkylating reagent for this step, chloromethyl ethyl
ether, is expected to be highly toxic. Current OSHA
guidelines indicate that chloromethyl methyl ether (a close
analogue to chloromethyl ethyl ether in structure and
reactivity) is a known human carcinogen. Therefore, we were
concerned about using chloromethyl ethyl ether as the
Results and Discussions
The aminolysis reaction to prepare penultimate intermedi-
ate 9 was typically carried out by treating ester intermediate
8 with excess 2-fluoroaniline in xylene at reflux temperature
for 4 h. After purification by chromatography, the desired
product was obtained in 58% yield. Due to the poor solubility
of the starting ester 8 as well as that of the carboxamide
product 9, this reaction was difficult to reproduce. Upon
scale-up, we encountered problems with inadequate stirring
of the heterogeneous reaction mixture and caking on the sides
of the reaction vessels. This led to difficulties with reaction
monitoring, and as a result, the isolated product was
contaminated with unreacted starting material and needed
further purification. We investigated several alternate solvents
for this step and found that a mixture of 5% (v/v) DMF/
xylene greatly improved the reaction. Ester 8 was more
soluble in this solvent combination; therefore, the reaction
(4) Buckley, J. Chem. Eng. News 1982, 60, 5.
(5) De Wall, G. Chem. Eng. News 1982, 60, 43.
Vol. 3, No. 4, 1999 / Organic Process Research & Development
•
261

Scheme 2
Scheme 3
Scheme 4
mixture was almost homogeneous at reflux temperature prior
to product precipitation. After 4 h, the reaction mixture was
filtered and washed with xylene to afford amide 9 in 93%
yield. Most importantly, the isolated product was obtained
in greater than 99% purity, thus eliminating any chromato-
graphic purification. Increasing the amount of DMF in the
solvent mixture did not afford any advantage; on the contrary,
the isolated yield was lower. The 5% DMF in xylene
conditions seemed to be optimum for this step (Scheme 2).
The final step of the synthesis is the ethoxymethylation
of 9 with chloromethyl ethyl ether in the presence of 15-
crown-5 to form RWJ-51204 (1). Due to the expected
toxicity and carcinogenicity of chloromethyl ethyl ether, we
investigated various approaches to replace it with a poten-
tially less toxic reagent. Our options were to identify an
alternate, less toxic alkylating reagent or to develop an in
situ preparation of chloromethyl ethyl ether to minimize the
occupational exposure risks during storage and handling. We
determined early on that 15-crown-5 was not necessary for
the alkylation, and thus, we eliminated the use of this toxic
reagent.
Diethoxymethane, a safer alternative to chloromethyl ethyl
ether, has been reported6,7 to be an effective ethoxymeth-
ylating reagent for alcohols, phenols, and amines in the
presence of acidic catalysts. Ozaki and co-workers6 reported
the alkylation of silylated fluorouracil derivatives using
diethoxymethane in the presence of Lewis acids such as
SnCl4. Similarly, Schaper7 reported alkylation of phenols
using diethoxymethane in the presence of a Montmorillonite
clay. However, all attempts at ethoxymethylation of inter-
mediate 9 under acidic conditions based on these reports
afforded none of the desired product, and only unreacted
starting material was recovered. We also investigated alky-
lations under basic conditions (stoichiometric or catalytic)
with various electrophiles including formaldehyde, dichlo-
romethane, bromochloromethane, or Mannich-type bases.
Again, we did not observe any product formation under these
conditions. Reasons for the failure of these reactions include
the limited solubility of the sodium salt of amide 9 in organic
solvents less polar than DMF and the poor nucleophilicity
of the carboxamide anion. In an effort to assess this, we
subjected the sodium salt of 9 to reaction with methyl iodide
(a very reactive alkylating reagent) in DMF. After 3 h at 60
°C in a sealed vessel only 40% conversion to the N-methyl
compound was observed. In light of the poor reactivity of 9
with such reactive alkylating reagents, more reactive elec-
trophilic ethoxymethylating reagents than those mentioned
above were designed and tested.
We proposed that oxonium intermediate 11 would be a
very reactive ethoxymethylating reagent. One way we
thought to prepare 11 was from quaternary ammonium salts
such as 10 (Scheme 3). We hypothesized that under basic
conditions 10 would convert to the reactive intermediate 11
upon elimination of N-methylpiperidine.
We prepared the quaternary ammonium iodide 10 by
treating N-(ethoxymethyl)piperidine with methyl iodide and
reacted it with the sodium salt of 9 in DMF (Scheme 4). At
ambient temperature, we detected no reaction; however upon
heating to reflux temperature for 2 h, we observed 50%
conversion to RWJ-51204 (1). We were concerned that at
elevated temperatures the alkylation proceeded through the
in situ formation of iodomethyl ethyl ether. This could occur
from nucleophilic attack by the iodide ion on the quaternary
ammonium salt and subsequent elimination of N-methyl-
piperidine. To test this hypothesis, we prepared the quater-
nary ammonium salt by treating N-(ethoxymethyl)piperidine
with methyl tosylate and repeated the ethoxymethylation
reaction. Even at elevated temperatures, we observed no
conversion to the desired product. This suggests that under
the reaction conditions, the quaternary ammonium salt does
not convert to oxonium intermediate 11 and that the active
alkylating reagent for the first reaction is iodomethyl ethyl
ether formed in situ under basic conditions.
Since we were unable to identify a less toxic ethoxy-
methylating reagent to carry out the alkylation, we investi-
gated the in situ formation of chloromethyl ethyl ether to
reduce the risk associated with exposure in handling and
storage of this reagent. On the basis of our previous results,
we examined the use of ethyl chloroformate to form
quaternary ammonium salt 12 which upon heating would
eliminate ethyl 1-piperidinecarboxylate to form chloromethyl
ethyl ether (Scheme 5). Treatment of N-(ethoxymethyl)-
piperidine with 1 equiv of ethyl chloroformate in DMF at
50 °C generated a mixture of the quaternary ammonium salt
12 and chloromethyl ethyl ether. However, after addition of
the sodium salt of carboxamide 9, the reaction only pro-
ceeded 60% (with the formation of many impurities) and
no further product was observed even upon prolonged
heating. This route would be problematic for large-scale
production for several reasons. The yield was low, and the
(6) Ozaki, S.; Watanabe, Y.; Fujisawa, H.; Hoshiko, T. Chem. Pharm. Bull.
1986, 34, 150.
(7) Schaper, U.-A. Synthesis 1981, 794.
262
•
Vol. 3, No. 4, 1999 / Organic Process Research & Development

Scheme 5
Table 1. Ethoxymethylation of 9 with alternate basesa
isolated product was contaminated with impurities and
unreacted starting material so that chromatographic purifica-
tion was still necessary. Also, N-(ethoxymethyl)piperidine
is not commercially available; therefore, we would need to
prepare it, thus adding another step to the synthesis.
Therefore, we investigated other in situ preparations of
chloromethyl ethyl ether.
RWJ-51204 RWJ-50171
CMEE
(equiv)
(1) area
(9) area
base
KOH
solvent
DMF
DMF
DMF
THF
DMF
THF/DMF
THF
DMF
DMF
% HPLC
% HPLC
1.5
2.0
1.0
1.0
1.0
1.1
1.7
1.7
2.3
2.3
0
38
0
23
55
75
6
82
100
100
100
56
100
77
45
25
94
18
0
NaOEt
Bu4NOH
Bu4NOH
KOt-Bu
KHMDS
DIPEA
DIPEA
DIPEA
In 1995, Bailey and co-workers8 reported the cleavage
of cyclic formaldehyde acetals with acetyl chloride in the
presence of Lewis acids to afford the corresponding chlo-
romethyl ethers which were then reacted with alcohols under
basic conditions to afford the desired products in high yield.
On the basis of this report, we investigated the formation of
chloromethyl ethyl ether from diethoxymethane and acetyl
chloride in the presence of catalytic sulfuric acid. A similar
approach was reported9 for the synthesis of chloromethyl
methyl ether from dimethoxymethane and acid chlorides in
methanol. Treatment of a slight excess (10 mol %) of
diethoxymethane with acetyl chloride and catalytic sulfuric
acid at 40 °C for 3 h resulted in complete conversion to
chloromethyl ethyl ether and ethyl acetate. Addition of the
sodium salt of 9 (generated from NaH in DMF) afforded
approximately 80% conversion to RWJ-51204 after 16 h at
ambient temperature. Addition of excess NaH completed the
reaction but led to dialkylated impurities confirmed by MS.
This was an important observation and the first time that
the reaction could be forced to completion. An aqueous
quench at the end of the reaction precipitated the product as
a tan solid in high yield. It also destroyed the residual
chloromethyl ethyl ether, thus eliminating the potential
exposure risk during workup. In order for this process to be
suitable for scale-up, we needed to replace the hazardous
NaH/DMF combination. Most importantly, to avoid any
chromatographic purification, we needed to completely
consume the difficult-to-remove starting material 9 without
forming any dialkylated impurities. Results of a survey of
DIPEA/DBU DMF
0
a Progress of reaction monitored by HPLC.
Scheme 6
alternate bases used for the N-ethoxymethylation of car-
boxamide 9 with chloromethyl ethyl ether are shown in Table
1.
As shown in Table 1, we achieved the best results using
diisopropylethylamine (DIPEA) as the base. Addition of
carboxamide 9 and DMF to an in situ generated solution of
chloromethyl ethyl ether followed by slow addition of DIPEA
(2.75 equiv) afforded clean conversion to RWJ-51204 (1).
Quenching with water after 16 h at ambient temperature
precipitated the crude product as an off-white solid in 85-
95% yield, > 99% purity.
Interestingly, we observed an initial rapid consumption
of the amide starting material 9 (∼90%) and conversion to
RWJ-51204 (1) in the first minutes after the reagents were
combined followed by an apparent slower but complete
disappearance of the remaining starting material over several
hours. Since we were concerned about the possible side
reaction between DIPEA and chloromethyl ethyl ether during
the prolonged reaction times, we examined this by 1H NMR
and observed a rapid reaction (t1/2 < 15 min) to form a new
(8) Bailey, W. F.; Zarcone, L. M. J.; Rivera, A. D. J. Org. Chem. 1995, 60,
2532.
(9) (a) Weinstock, L. M.; Karady, S.; Sletzinger, M. U.S. Patent 3,972,947,
1975. (b) Amato, J. S.; Karady, S.; Sletzinger, M.; Weinstock, L. M.
Synthesis 1979, 970. (c) Linderman, R. J.; Jaber, M.; Griedel, B. D. J. Org.
Chem. 1994, 59, 6499.
Vol. 3, No. 4, 1999 / Organic Process Research & Development
•
263

Scheme 7
U
species, presumably the quaternary ammonium salt 13
(Scheme 6). Our current hypothesis is that the reaction
between these two reagents (DIPEA and CMEE) is fast but
reversible and that the quaternary ammonium salt 13 serves
as a source of chloromethyl ethyl ether to complete the
reaction in the presence of excess base. This is consistent
with our partial successes in generating chloromethyl ethyl
ether from other aminomethyl ethyl ethers on quaternization.
In preliminary studies on recrystallization of crude RWJ-
51204, we found that isopropyl alcohol improved the color
and provided good recovery of the solid with high through-
put. We were concerned about possible loss of the ethoxy-
methyl side chain thermally in alcoholic solvent; however,
we found no evidence of dealkylation or decomposition of
product after 17 h at reflux in isopropyl alcohol.
For the final process, there was only one outstanding issue,
the poor solubility of intermediate 9 in DMF. After we
prepared the chloromethyl ethyl ether, the reaction flask had
to be opened and the amide added as a solid. To avoid the
risk associated with possible exposure to chloromethyl ethyl
ether upon opening the reaction vessel, we investigated
alternate methods of addition. Our first approach involved
in situ formation of chloromethyl ethyl ether in the presence
of amide intermediate 9 which upon addition of DIPEA
would afford RWJ-51204 in a “one-pot” reaction. However,
once all the reagents were combined in the presence of
catalytic acid, the reaction mixture became very thick and
difficult to stir. Since this would be problematic for large-
scale production, we examined formation of the anion of
carboxamide 9 with various bases (other than DIPEA) in
DMF in hopes of improving solubility. After investigating
several bases, we found that treatment of intermediate 9 with
1 equiv of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in
DMF yielded a homogeneous solution. Addition of this
homogeneous solution to a closed reaction vessel containing
chloromethyl ethyl ether under inert conditions followed by
slow addition of excess DIPEA afforded clean conversion
to RWJ-51204 (1). Quenching with water precipitated the
product as an off-white solid in 95% yield. The crude product
was recrystallized from isopropyl alcohol to give RWJ-51204
(1) in 92% yield for the two steps and >99% purity (Scheme
7).
In summary, we have successfully developed a safe large-
scale synthesis of RWJ-51204 which avoids the safety
challenges of the original medicinal chemistry route. We
increased the yield and throughput for the aminolysis step
to prepare carboxamide 9 by addition of 5% DMF to the
reaction mixture. We improved the final alkylation step by
developing an in situ preparation of chloromethyl ethyl ether
to minimize the potential exposure risks during storage and
handling. We also eliminated the use of 15-crown-5 and
replaced the hazardous combination of NaH/DMF with
DIPEA/DBU/DMF. As a result, we increased the yield for
the final step after purification from 60% to 92%, while
eliminating chromatography. With these improvements, we
prepared RWJ-51204 (1) in 36% overall yield and applied
this process to synthesize drug substance for use in early
development toxicological and clinical studies.
Experimental Section
General Procedures. Melting points were determined
with a Thomas-Hoover capillary melting point apparatus and
are uncorrected. 1H NMR spectra were recorded on a Bruker
AM 300, 300 MHz spectrometer. The chemical shifts were
expressed as δ units with Me4Si as the internal standard
1
(multiplicities in H NMR are referred to as follows: s for
singlets, d for doublets, t for triplets, q for quartets, and m
264
•
Vol. 3, No. 4, 1999 / Organic Process Research & Development

for multiplets). All reactions are carried out under a nitrogen
or an argon atmosphere. Solvents and reagents were obtained
from commercial sources and used without further treatment
or purification.
lamine (22.6 g, 0.17 mol) was added, and the reaction
mixture was stirred for 3 h at ambient temperature. The
reaction mixture was cooled in a water bath, and water (60
g) was added slowly. The resulting suspension was stirred
for 30 min. The solid was collected by filtration and washed
with water (250 g). The solid was dried in vacuo at ambient
temperature to afford the desired crude product 1 (38.0 g,
95.0%) as a light yellow, granular, free-flowing powder.
The crude product was dissolved in isopropyl alcohol (196
g) and heated to reflux. The hot solution was filtered to
remove particulates and washed with hot isopropyl alcohol
(8 g). The filtrate was concentrated to remove isopropyl
alcohol (90 g) via distillation. The clear solution was cooled
to ambient temperature slowly, resulting in crystallization.
The solid was collected by filtration, washed with isopropyl
alcohol (50 g), and then dried in vacuo at 25 °C to afford 1
Pyrido[1,2-a]benzimidazole-4-carboxamide, 7-Fluoro-
N-(2-fluorophenyl)-1,2,3,5-tetrahydro-3-oxo (9). To a stirred
suspension of ethyl pyrido[1,2-a]benzimidazole-4-carboxy-
late, 7-fluoro-1,2-dihydro-3-oxo (8)2 (27.6 g, 0.10 mol) in
DMF (22 g) and xylene (384 g) under a nitrogen atmosphere
was added 2-fluoroaniline (36.0 g, 0.32 mol), and the
suspension was brought to a gentle reflux. The reaction
mixture was stirred at reflux for 4 h and then cooled to 20
°C. The solid was collected by filtration and washed with
xylene (70 g) and then air-dried to yield 9 (31.58 g, 92.6%)
as a tan solid: mp ) 266-268 °C; 1H NMR (DMSO-d6) δ
(ppm) 2.79 (2H, t), 4.31 (2H, t), 7.00 (1H, m), 7.12-7.29
(3H, m), 7.41 (1H, dd), 7.57 (1H, m), 8.50 (1H, t), 12.18
(1H, s, NH), 12.72 (1H, s, NH). MS (ESI) m/z 342 (MH+).
Anal. Calcd for C18H13F2N3O2: C, 63.34; H, 3.84; N, 12.31.
Found: C, 63.29; H, 3.81; N, 12.35.
Pyrido[1,2-a]benzimidazole-4-carboxamide, 5-(Ethoxy-
methyl)-7-fluoro-N-(2-fluorophenyl)-1,2,3, 5-tetrahydro-
3-oxo (1). To a stirred solution of diethoxymethane (28.6 g,
0.27 mol) and acetyl chloride (19.6 g, 0.25 mol) under a
nitrogen atmosphere was added sulfuric acid (0.30 g, 0.003
mol). The resulting clear solution was warmed to 40 °C for
3 h, and the reaction mixture was then cooled to ambient
temperature. A solution of pyrido[1,2-a]benzimidazole-4-
carboxamide, 7-fluoro-N-(2-fluorophenyl)-1,2,3,5-tetrahydro-
3-oxo (9) (34.1 g, 0.10 mol) and 1,8-diazabicyclo[5.4.0]-
undec-7-ene (15.2 g, 0.10 mol) in DMF (97.0 g) was
prepared and added to the reaction mixture slowly to maintain
the temperature between 25 and 35 °C. Diisopropylethy-
1
(35.7 g, 94%) as a tan solid: mp 153-155 °C; H NMR
(CDCl3) δ (ppm) 1.08 (3H, t), 2.85 (2H, t), 3.33 (2H, q),
4.15 (2H, t), 5.75 (2H, s), 6.97 (1H, m), 7.10-7.13 (1H,
m), 7.22 (1H, m), 7.35 (1H, dd), 8.38 (1H, t), 12.01 (1H, s).
MS (ESI) m/z 400 (MH+). Anal. Calcd for C21H19F2N3O3:
C, 63.15; H, 4.79; N, 10.52. Found: C, 63.03; H, 4.66; N,
10.40.
Acknowledgment
We thank Donna Heidel, Staff Industrial Hygiene Spe-
cialist, for her assistance with the occupational health/safety
aspects of this work. We also thank Sergio Cesco-Cancian
and John T. Hortenstine for the scale-up work and Robin R.
Henneman-Webster for helpful discussions.
Received for review February 10, 1999.
OP990182L
Vol. 3, No. 4, 1999 / Organic Process Research & Development
•
265