The process development of a novel aldose reductase inhibitor, FK366. Part 1. Improvement of discovery process and new syntheses of 1-substituted quinazolinediones

Article abstract of DOI:10.1021/op0340661

This contribution describes part 1 of process development of a novel aldose reductase inhibitor FK366 (1). The original process applied on a laboratory scale was improved from the safety viewpoint to manufacture materials on 500-L scale suitable for toxic

Full text of DOI:10.1021/op0340661

Organic Process Research & Development 2003, 7, 700−706
The Process Development of a Novel Aldose Reductase Inhibitor, FK366. Part
1. Improvement of Discovery Process and New Syntheses of 1-Substituted
Quinazolinediones
Shunsuke Goto,* Hiroyuki Tsuboi, Masami Kanoda, Koji Mukai, and Kooji Kagara
Chemical DeVelopment Laboratories, Fujisawa Pharmaceutical Industry Co. Ltd., 2-1-6,
Kashima, Yodogawa-ku, Osaka 532-8514, Japan
Abstract:
This contribution describes part 1 of process development of a
novel aldose reductase inhibitor FK366 (1). The original process
applied on a laboratory scale was improved from the safety
viewpoint to manufacture materials on 500-L scale suitable for
toxicological and pharmacological evaluations. A new process,
including regioselective alkylation of silylated quinazolinedione,
provided a practical and cost-effective synthesis of FK366 in a
dramatically increased yield.
Introduction
Figure 1. Aldose reductase inhibitors.
Diabetic complications represented by neuropathy, cata-
ract, and retinopathy are brought about by the overaccumu-
lation of sorbitol in tissues. Aldose reductase inhibitors
(ARIs) which inhibit an enzyme of the polyol pathway were
explored for the remedy of diabetic complications (Figure
1).1 3-(4-Bromo-2-fluorobenzyl)-7-chloro-3,4-dihydro-2,4-
dioxo-1(2H)-quinazolineacetic acid (FK366, 1),2 a potent
ARI, was invented at Fujisawa laboratories.
The original process using an isatoic anhydride involved
several problems in scale-up, and we improved it to the
procedure applicable to pilot manufacturing of several 10
kilograms of drug substance. Whilst the improved eight-step
synthesis was efficient enough to give a material suitable
for toxicological studies on early stages, a more cost-effective
process was required. During further studies we discovered
that bis-silylated-2,4-quinazolinedione reacts with alkyl
halides to give 1-substituted quinazolinediones exclusively.3
We applied this reaction to the manufacture of FK366 and
developed a straightforward five-step synthesis suitable for
the production of FK366 on a commercial scale.
but also for protection (during N-alkylation) and activation
(for amide formation) of carboxylic acid. However, the
transformation of 3a to 4 afforded harmful benzylbromide
as a byproduct.
The N-alkylation of 4 employed sodium hydride in
dimethylformamide (DMF) system, which had a potential
hazard for a large-scale synthesis.4 The cyclization of
anthranilamide (5) to quinazolinedione (6) with carbonyl-
diimidazole (CDI) included another problem. A mixture of
5 and CDI in dioxane was heated, followed by distillation
of dioxane under atmospheric pressure. A cyclization reaction
suddenly occurred at the end of this procedure, and the
reaction mixture solidified, accompanying severe exotherm.
In the case where the reaction was conducted in the presence
of solvent, the cyclization reaction proceeded sluggishly. The
N3 benzylation of quinazolinedione nucleus also utilized
sodium hydride in DMF.
Process Improvements of Original Process. Considering
the mechanism of isatoic anhydride ring closure, the Cbz
group of 3a was substituted to ethoxycarbonyl group of 3b,
and the byproduct involved was the less harmful ethyl
bromide (Scheme 2).
The investigations in the cyclization reaction indicated
that the reaction was self-accelerated by imidazole liberated
from CDI during the reaction. The reaction carried out in
dioxane as a solvent in the presence of a catalytic amount
of imidazole was easier to control.
Results and Discussion
Original Process for FK366. As for the introduction of
carboxymethyl side chain to the quinazolinedione nucleus
of FK366, the original route (Scheme 1) exploited an isatoic
anhydride (4) derived from N-Cbz-anthranilic acid (3a). 4
was a useful intermediate not only for selective N-alkylation
* Author for correspondence: E-mail shunsuke_goto@po.fujisawa.co.jp.
Telephone +81-6-6390-1202. Fax +81-6-6390-4419.
(1) (a) Humber, L. G. In Progress in Medicinal Chemistry; Ellis, G. P., West,
G. B., Eds.; Elsevier Science Publishers: Amsterdam, 1987; Vol. 24, pp
299. (b) Malamas, M. S.; Millen, J. J. Med. Chem. 1991, 34, 1492-1503.
(2) (a) Sei, S.; Nozu, K. Diabetes 1989, 32 (Suppl. 1), 122. (b) Ao, S.; Kikuchi,
C.; Notsu, Y.; Yamaguchi, I. Jpn. J. Pharmacol. 1989, 49 (Suppl.).
(3) Goto, S.; Tsuboi, H.; Kagara, K. Chem. Express 1993, 8, 761-764.
For a large-scale synthesis, powdered potassium carbon-
ate5 in DMF was found to be a suitable alternative to the
(4) Laird, T. Chem. Ind. 1986, 17, 134.
(5) Commercially available from Nippon Soda Co., Ltd.
700
•
Vol. 7, No. 5, 2003 / Organic Process Research & Development
10.1021/op0340661 CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/20/2003

Scheme 1. Original Process for 1
Scheme 2. Ring closure to isatoic anhydride (4)
Scheme 3. Retrosynthesis of FK366 (1)
Scheme 4. Direct alkylation of quinazolinedione (9)
potentially hazardous system of sodium hydride in DMF for
the introduction of side chains.
Together with these major process improvements, each
step was reexamined from the viewpoints of safety, yield,
and operability. The improved process was successfully
scaled up to 500 L, manufacturing 24 kg of drug substance
suitable for toxicological studies. Overall yield from 2-amino-
4-chlorobenzoic acid (2) to 1 was improved from 27% to
65%.
Table 1. Alkylation of 9 with ethylbromoacetate in the
Development of Silylquinazoline Method. Although the
improved original route was efficient enough to manufacture
several tens of kilograms of drug substance, it consisted of
eight manufacturing steps including purification of interme-
diates and final product. A more efficient and inexpensive
process was still required for further stages.
As depicted in the retrosynthetic analysis, we envisaged
a new method for selective introduction of alkyl chains to
quinazolinedione (9), which was easily obtained from 2. The
conceptually simple way to FK366 was selective introduction
of a carboxymethyl group to N1 position followed by a
halobenzyl group to N3 position (Scheme 3).
presence of K2CO3 in DMF
compound
9
6
10
11
peak area (%)
37.1
18.7
3.8
38.7
improve selectivity with the combinations of different bases,
solvents, and temperature were in vain (Scheme 4, Table
1).
N1-selective alkylation of pyrimidinedione6 is achieved
via bis(trimethylsilyloxy)-pyrimidine intermediate which is
commonly exploited for the synthesis of pyrimidine nucle-
(6) (a) Sakai, T. T.; Pogolotti, A. L.; Santi, D. V. J. Heterocycl. Chem. 1968,
5, 849. (b) Micklitz, W.; Lippert, B.; Schllhorn, H.; Thewalt, T. U. J.
Heterocycl. Chem. 1989, 26, 1499. (c) Singh, H.; Aggarwal, P.; Kumar, S.
Synthesis 1990, 520.
The reaction of 9 and ethyl bromoacetate with potassium
carbonate as a base in DMF afforded a mixture of 9, mono-
(6 and 10) and dialkylated products (11). The efforts to
Vol. 7, No. 5, 2003 / Organic Process Research & Development
•
701

Scheme 5. Silylquinazoline method for 6
otides.7 This procedure is also applied to the preparation of
quinazoline nucleosides.8 Thus, we applied these methodolo-
gies for the regioselective alkylation of quinazolinediones.
We found that if bis(trimethylsilyloxy)quinazoline (12)
derived from 9 was heated in ethyl bromoacetate, the desired
1-ethoxycarbonylmethyl derivative 6 was selectively obtained
without involving other byproducts. The following introduc-
tion of 2-fluoro-4-bromobenzyl moiety at the N3 position
using the corresponding benzylbromide and K2CO3 as a base
afforded the penultimate ethyl ester (8) in good yield
(Scheme 5).
Figure 2. Ureide formation and pH in the reaction mixture
(HCl added in two portions).
from 3 to 1.5 equiv, and the yield was improved from 85 to
94% on 500-L scale (Scheme 7, Figure 2, Figure 3).10
Silylation of Quinazolinedione (9). The silylating agent
was selected from the viewpoints of convenience and
economy. Silylation with trimethylchlorosilane and triethyl-
amine in dichloromethane proceeded well, but the trouble-
some filtration of triethylamine hydrochloride was necessary.
We used hexamethyldisilazane (HMDS) as a clean silylating
agent and ammonium sulfate as a silylation catalyst. In early
studies, the reaction speed was not consistent. For example,
several hours were needed until the reaction was completed
on a large scale, whilst an hour was needed on a small scale.
We speculated that the removal of ammonia from the reaction
mixture might be crucial for the silylation with HMDS, and
thus the agitation and reflux speed might be influential to
the silylation period. Figures 4 and 5 show the correlation
between silylation time and agitation speed or bath temper-
ature at scale-down experiments on 500-mL scale using
geometrically similar equipment. With stronger agitation and
higher bath temperature, the silylation smoothly proceeds
to give a clear solution.
Thus, we developed a new silylquinazoline process for
FK366 (Scheme 6).
Scale-Up Study of Silylquinazoline Process: Synthesis
of 7-Chloroquinazoline-2,4-dione (9). The well-known
procedure to manufacture quinazolinedione from anthranilic
acid described in “Organic Synthesis”9 could not be directly
applied for the synthesis of 4-chloroanthranilic acid since
the initial urea formation reaction did not proceed due to
the exceptionally low solubility in the reaction medium. After
dissolving 4-chloroanthranilic acid as a sodium salt in water,
potassium cyanate was added, and then diluted HCl was
added in two portions to promote the ureide-forming reaction.
To the obtained solution of 2-ureide benzoic acid was added
sodium hydroxide for the cyclization to afford 7-chloro-
quinazolinedione sodium salt in good yield. We found that
the ureide formation was highly dependent on the pH of the
reaction mixture. Whilst the reaction proceeded sluggishly
at the pH of >7, the decomposition of cyanate was
predominant at the pH of <5. As a result, the reaction was
susceptible to the mixing conditions and was not reproducible
in the yield. Thus, the best result of the reaction in the yield
and reproducibility was realized by continuously adjusting
pH to 6-7 with HCl. As a result of this pH-controlling
procedure, the amount of potassium cyanate was reduced
Next, we investigated more effective catalysts for the
silylation. Pulverized ammonium sulfate (22 µm) was a better
catalyst than a regular granule (350 µm), and to our surprise,
sulfuric acid and methanesulfonic acid were highly efficient
catalysts (Table 2).
Scheme 6. Silylquinazoline process for FK366 (1)
702
•
Vol. 7, No. 5, 2003 / Organic Process Research & Development

Figure 5. Effect of bath temperature on silylation time. A
mixture of 9 (20 g), HMDS (2 mol equiv) and (NH4)2SO4 (350
µm, 0.15 mol equiv) were refluxed in toluene (60 mL). Agitation
speed ) 400 rpm. The period to obtain a clear solution was
regarded as a silylation time.
Figure 3. Ureide formation and pH in the reaction mixture
(pH controlled to 6-7).
Table 2. Effects of catalysts for silylationa
amount of catalyst
(equiv)
silylation time
(min)
catalyst
(NH4)2SO4 (350 µm)
(NH4)2SO4 (22 µm)
p-TsONH4
H2SO4
H2SO4
0.15
0.15
0.15
0.15
0.08
0.08
90
70
90
50
50
45
MsOH
a A mixture of 9 (20 g), HMDS (2 mol equiv), and catalyst were refluxed in
toluene (60 mL) at bath temperature 135 °C.
Caution! As concentrated sulfuric acid and HMDS react
Violently, it must be added with greatest care on a large
scale.
Figure 4. Effect of agitation speed on silylation time. A mixture
of 9 (20 g), HMDS (2 mol equiv) and (NH4)2SO4 (350 µm, 0.15
mol equiv) were refluxed in toluene (60 mL). Bath temperature
) 135 °C. The period to obtain a clear solution was regarded
as a silylation time.
Alkylation of Bis-trimethysilyloxyquinazoline. Regio-
selective introduction of ethoxycarbonylmethyl group at the
N1 position of the quinazolinedione nucleus was achieved
using ethyl bromoacetate as a reagent/solvent. After the
silylation with HMDS in toluene, the reaction mixture was
concentrated under reduced pressure to give bis-trimethysilyl
compound as a ceraceous solid, which was heated in the
presence of 2 volumes of ethyl bromoacetate above 120 °C
for several hours. To the reaction mixture was added dioxane
to improve the agitation, and then methanol was added
dropwise to quench the monosilylated intermediate, giving
a slurry of the desired product 6. Methanol also played a
role of decomposing trimethylsilyl bromide, a byproduct of
alkylation.
Scheme 7. pH dependence of ureide formation
Whilst the granular ammonium sulfate tended to block
up a bottom discharge valve of the reactor and to be
contaminated in the product, the catalyst prepared in
situ from sulfuric acid and HMDS did not involve such
problems.
The reaction in the presence of other solvents such as
toluene, DMF, or diglyme with ethyl bromoacetate to reduce
the amount of this harmful reagent did not give the product
in the acceptable yield since the reaction was retarded. The
substitution of expensive and highly lachrymatory ethyl
bromoacetate for ethyl chloroacetate was in vain since over
100 hours were required before the reaction led to completion
at the same temperature.
(7) A review for the silyl method of synthesis of nucleosides and nucleotides.:
Lukevics, E.; Zablotskaya, A. E.; Solomennikova, I. I. Russ. Chem. ReV.
1974, 43, 140.
(8) Stout, M. G.; Robins, R. K. J. Org. Chem., 1968, 33, 1219.
(9) Lange, N. A.; Sheibley, F. E. Org. Synth. 1943, 2, 79-80.
(10) Similar results were reported by P. D. Hammen et al.: Hammen, P. D.;
Allen, D. J. M. J. Heterocycl. Chem. 1987, 24, 1701-1703.
Benzylation at the N3 Position. In the early process,
N-benzylation was conducted using the NaH/DMF system.
To avoid this potentially hazardous system, we explored other
Vol. 7, No. 5, 2003 / Organic Process Research & Development
•
703

Figure 6. XRPD pattern of FK366: Form A (left), Form B (right).
Scheme 8. Origin of impurity 13
reaction conditions. Further investigations indicated that
powdered potassium carbonate in acetone was an excellent
alternative for this reaction. 6 in acetone was then refluxed
with 4-bromo-2-fluorobenzylbromide (7) in the presence of
powdered K2CO3 as a base5 for several hours to give 8 in
high yield. After the completion of the reaction, acetic acid
was added to decompose excess potassium carbonate, and
then the mixture was refluxed again to decompose excess
benzylbromide by the reaction with potassium acetate to
avoid hazardous exposure to personnel during workup and
filtration. To the reaction mixture was added methanol11 to
improve stirrability, and crystals were collected, followed
by drying under reduced pressure to give 8.
Table 3. Effect of spiked amount of 13 and seed crystal
toward crystal form of FK366
seed crystal
spiked amount (wt %)
crystal form obtained
none
0.1
0.5
1.0
A
A
B
Hydrolysis and Purification. Hydrolysis of ethyl ester
was conducted with sodium hydroxide in aqueous methanol.
After the completion of the reaction, the reaction mixture
was acidified with concentrated hydrochloric acid (HCl) to
precipitate crude FK366. Purified FK366 was obtained by
recrystallization of crude material in aqueous 2-propanol.
After dissolution in hot aqueous 2-propanol, the product was
treated by charcoal and then filtered. To the filtrate was added
purified water, and then the solution was cooled to give
purified FK366 as white crystals.
The corresponding methyl and isopropyl esters were major
impurities in FK366 drug substance involved in these steps.
The amount of HCl, the water-solvent ratio, and crystal-
lization temperature were studied to reduce these byproducts.
Emergence of a New Polymorph. In early studies, only
one crystal form (A) was identified for FK366. However a
new crystal form (B) was obtained in the course of further
development of new silylquinazoline process (Figure 6), and
precipitation of the B-form crystal was problematic on a
large-scale synthesis due to highly thick slurry and poor
filterability.
A
B
0
0
A
A
A
B
0.5
0.5
A
B
A
B
2.0
2.0
B
B
degree. Table 3 shows the correlation of crystal form with
the spiked amount of 13. Addition of 1% of 13 resulted in
precipitation of the B-form crystal in the absence of seed
crystal. On the other hand, A-form material was always
obtained independently from the seed crystal without 13.
With 0.5% of 13, the crystal form depended on the seed
crystal form, and with 2% of 13, the B-form was always
obtained independently from the seed form.12
There is a high possibility that this impurity was involved
in the new process, and thus we endeavored to control the
amount of dibenzyl compound. To control the contamination
of this impurity below 1%, the alkylation reaction was
monitored by HPLC, confirming the residual amount of
unreacted 9 was below 1% as well.
Selective Monoalkylation of Quinazolinediones. We
examined the synthetic applicability of selective N1 alky-
lation of quinazolinedione via bis-trimethylsilylated inter-
mediates. As shown in Table 4, several 1-substituted
quinazolinediones were obtained in high yields. We believe
the reaction is a convenient general method to manufacture
N1 substituted quinazolinediones.3
The crystal form was difficult to control in early studies
since the form varied from lot to lot and seemed independent
from the quality of the product. However, detailed investiga-
tions into impurities included in FK366 revealed that N,N′-
di-benzyl compound (13), which had originated from the bis-
benzylation of incorporated intermediate 9 in 6, played an
important role for the crystallization of the polymorph
(Scheme 8).
Recrystallization of FK366 with a spiked amount of 0.5-
1% of 13 led to consistent precipitation of the B-form crystal
even though original FK366 never arose as B-form to any
(12) Influences of impurities or additives to the plymorph control were reported
in the following literature: (a) Ebian, A. R.; Moustafa, M. A.; Khalil, S.
A.; Motawi, M. M. J. Pharm. Pharmacol. 1973, 25, 13-20. (b) Hasegawa,
M.; Fukuda, N.; Higuchi, H.; Noguchi, S.; Matsubara, I. Agric. Biol. Chem.
1977, 41, 49. (c) Kitamura, M.; Funahara, H. J. Chem. Eng. Jpn. 1994, 27,
124-126. (d) Davey, R. J.; Blagden, N.; Potts, G. D.; Docherty, R. J. Am.
Chem. Soc. 1997, 119, 1767-1772.
(11) If the methanol was added without decomposing excess potassium carbonate,
a considerable amount of methylester was contaminated.
704
•
Vol. 7, No. 5, 2003 / Organic Process Research & Development

Table 4. Selective N1 alkylation of quinazolinediones
pH of the suspension was adjusted to 1.0-1.5 with 35% HCl
at 30 ( 2 °C, and stirring was continued 2 h. After cooling
below 5 °C, the precipitates were collected by a centrifuge,
washed with water (40 L), and dried in a vacuum tray dryer
to give 9 (19.4 kg, 94.0% yield) as a white powder: mp 231-
233 °C.
1H NMR (200 MHz, DMSO-d6, δ) 7.17 (1H, d, J ) 1.8
Hz), 7.22 (1H, dd, J ) 8.3, 1.8 Hz), 7.88 (1H, d, J ) 8.3
Hz), 11.26 (1H, brs), 11.42 (1H, brs); IR (ATR) 3042, 2841,
1679, 1617, 1428, 1285, 1081 cm-1; MS (EI) m/z 196 (M+).
Anal. Calcd for C8H5N2O2Cl: C, 48.88; H, 2.56; N, 14.25.
Found: C, 48.89; H, 2.59; N, 14.14.
X
R
yield(%)
7-Cl
H
7-Cl
H
H
H
CH2CO2Et
CH2CO2Et
Et
97.7
97.4
79.2
84.4
96.1
77.4
Et
CH2Ph
(CH2)13CH3
Ethyl 7-Chloro-3,4-dihydro-2,4-dioxo-1(2H)-quinazo-
lineacetate (6)
To a suspension of 9 (18.25kg) in toluene (54.8L) and
hexamethyldisilazane (HMDS; 36.0kg) was added sulfuric
acid (0.73kg) with greatest cautions. The mixture was heated
to reflux and stirred under heavy refluxing for 8 h till clear
solution was obtained. After the removal of toluene and
excess HMDS with vacuum distillation, ethyl bromoacetate
(36.5L) was added to the residue. The reaction mixture was
heated to 115-130 °C and was stirred at this temperature
for 3 h. After checking the completion of the reaction with
HPLC, the reaction mixture was diluted with 1,4-dioxane
(36.5 L) at 100 °C, and then methanol (54.8 L) was added
at 70 °C for 30 min.
The suspension was cooled below 5 °C and precipitates
were collected by a centrifuge. After washing with methanol
(20 L) and water (90 L), the wet crystals were dried in a
vacuum tray dryer to afford 6 (25.4kg, 96.8%) as a white
powder: mp 253-255 °C.
Conclusions
A concise five-step sequence has been developed and
demonstrated for the preparation of FK366 on 20-30 kg
scale in 80% overall yield. This new process is characterized
by the selective monoalkylation of the quinazolinedione
nucleus and has led to dramatic improvement in overall yield
in comparison to that of the original method (27%). The
reported methods are practical, inexpensive, and suitable for
a large-scale synthesis.
Experimental Section
Solvents and reagents were obtained from commercial
sources and were used without any purification. The melting
points were uncorrected. IR spectra were recorded on a
SensIR Technologies TravelIR spectrometer. 1H NMR
spectra were obtained on a Brucker DPX200 using tetram-
ethylsilane as internal standard. Mass spectra were recorded
on a Hewlett-Packard 1100LC/MSD mass spectrometer using
EI for ionization. Elemental analyses were carried out on a
Perkin-Elmer CHN elemental analyzer. HPLC analyses were
performed using a YMC gel ODS 120 Å S-7 column and a
CH3CN- or CH3OH-buffer phase. Reactions were carried out
in 200-1000-L glass-lined carbon steel reactors. The inter-
mediates and the drug substance were isolated in a stainless
steel centrifuge (30-in. i.d.) using a polypropylene filter bag
and dried in a vacuum tray dryer.
7-Chloro-3,4-dihydro-2,4-dioxo-1(2H)-quinazoline (9).
To a suspension of 2-amino-4-chlorobenzoic acid 2 (18.0
kg) in water (100 L) was added aqueous sodium hydroxide
solution (4.6 kg in 20 L) to give a clear solution. To the
reaction mixture was added aqueous sodium cyanate solution
(11.8 kg in 240 L). A process pH electrode was immersed
into the vessel, and the pH value of the solution was adjusted
to 6.5 with 35% HCl at room temperature to give a porridgy
suspension. The pH of the reaction mixture was maintained
at 6.5 ( 0.2 with 35% HCl for 5 h with stirring at 30 ( 2
°C. After the completion of the ureide formation, to the
reaction mixture was added sodium hydroxide (16.8 kg). The
reaction mixture was heated to 60-65 °C and stirred for 2
h at this temperature (sodium salt of quinazoline dione
precipitates). After cooling below 5 °C, the precipitates were
collected by a centrifuge to give sodium salt as a white mass.
Sodium salt thus obtained was returned to the vessel and
was suspended in acetone (180 L) and water (180 L). The
1H NMR (200 MHz, DMSO-d6, δ) 1.22 (3H, t, J ) 7.2
Hz), 4.17 (2H, q, J ) 7.2 Hz), 4.91 (2H, s), 7.35 (1H, dd, J
) 8.5, 1.6 Hz), 7.61 (1H, d, J ) 1.6 Hz), 8.01 (1H, d, J )
8.5 Hz), 11.89 (1H, br s); IR (ATR) 3042, 1737, 1679, 1366,
1081, 842 cm-1; MS (EI) m/z 282 (M+). Anal. Calcd for
C12H11N2O4Cl: C,50.99; H, 3.92; N, 9.91. Found: C, 50.94;
H, 3.86; N, 9.83.
Ethyl 3-(4-Bromo-2-fluorobenzyl)-7-chloro-3,4-dihy-
dro-2,4-dioxo-1(2H)-quinazoline Acetate (8). To a suspen-
sion of 6 (25.0 kg) in acetone (125 L) were added pulverized
potassium carbonate (14.7 kg) and 4-bromo-2-fluorobenzyl-
bromide (30.8 kg) at ambient temperature. The reaction
mixture was heated to reflux and stirred for 5 h.
After checking the end of the alkylation, acetic acid (12.5
kg) was added at 40-45 °C, and stirring was continued for
2 h under reflux to decompose excess benzylbromide. To
the reaction mixture was added methanol (200 L) at 55-60
°C for 30 min, and then it was cooled below 5 °C.
Precipitates were collected by centrifuge and washed with
methanol (50 L) and water (250 L). The wet crystals were
dried in a vacuum tray drier to afford 8 (39.5kg, 95.1%) as
white crystals: mp 160-161 °C.
1H NMR (200 MHz, DMSO-d6, δ) 1.20 (3H, t, J ) 7.1
Hz), 4.17 (2H, q, J ) 7.1 Hz), 4.99 (2H, s), 5.13 (2H, s),
7.15 (1H, t, J ) 8.3 Hz), 7.32-7.42 (2H, m), 7.55 (1H, dd,
J ) 8.5, 1.6 Hz), 7.71 (1H, d, J ) 1.6 Hz), 8.08 (1H, d, J )
8.5 Hz); IR (ATR) 3042, 2841, 1737, 1679, 1617, 1366, 1227
Vol. 7, No. 5, 2003 / Organic Process Research & Development
•
705

cm-1; MS (EI) m/z 492 (M + 23). Anal. Calcd for
C19H15N2O4BrClF: C, 48.59; H, 3.22; N, 5.96. Found: C,
48.70; H, 3.16; N, 5.88.
The precipitates were collected by centrifuge, washed with
aqueous 2-propanol (2-propanol 36 L + purified water 36
L), and dried in a vacuum tray dryer to give 1 (34.7kg,
96.4%) as a white crystalline powder: mp 222.5-223 °C.
1H NMR (200 MHz, DMSO-d6, δ) 4.91 (2H, s), 5.14 (2H,
s), 7.15 (1H, t, J ) 8.3 Hz), 7.31-7.41 (2H, m), 7.55 (1H,
dd, J ) 8.5, 1.6 Hz), 7.67 (1H, d, J ) 1.6 Hz), 8.08 (1H, d,
J ) 8.5 Hz), 13.28 (1H, br s); IR (ATR) 3042, 2841, 1737,
1679, 1617, 1594, 1428, 1227 cm-1; MS (EI) m/z 441 (M+).
Anal. Calcd for C17H11N2O4BrClF: C, 46.23; H, 2.51; N,
6.34. Found: C, 46.24; H, 2.47; N, 6.32.
General Procedures for the Selective Alkylation of
Quinazolinedione. To a suspension of quinazolinedione in
toluene and HMDS was added sulfuric acid at room
temperature. The reaction mixture was refluxed till the clear
solution was obtained. The solution was concentrated under
vacuum to give a ceraceous solid. The residue was combined
with an alkyl halide, leading to alkylation under elevated
temperature. After the completion of the reaction, methanol
was added to the reaction mixture to crystallize 1-substituted
quinazolinedione.
3-(4-Bromo-2-fluorobenzyl)-7-chloro-3, 4-dihydro-2,
4-dioxo-1(2H)-quinazolineacetic Acid (1). To the suspen-
sion of 8 (39.0kg) in methanol (546 L) was added aqueous
sodium hydroxide solution (4.98 kg in 156 L) at 30-40 °C.
The reaction mixture was heated to reflux and stirred for 45
min until the completion of the reaction. The hot solution
was filtered to remove insoluble particulate, and the vessel,
the filter, and the connecting piping were rinsed with aqueous
methanol (methanol 78 L + water 78 L). The filtrate and
the rinse solution were combined and diluted with water (78
L). The solution was cooled to 30 °C, and 35% HCl (19.5L)
was added for the crystallization at 30-35 °C. After stirring
for 30 min at this temperature, the suspension was cooled
below 5 °C. The precipitates were collected by centrifuge
and washed with aqueous methanol (methanol 117 L + water
117 L). The wet crystals were dried in a vacuum tray dryer
to afford crude 1 (36.15kg, 98.6%) as a white powder.
The crude crystals of FK366 (1) thus obtained (36.0 kg)
were suspended in aqueous 2-propanol (2-propanol 360 L
+ purified water 72 L). The suspension was heated to reflux
to give a clear solution and was treated with charcoal (1.08
kg) for 30 min. The charcoal was filtered off, and the filter
cake was washed with 2-propanol (72 L). To the filtrate was
added purified water (72 L) at 75-80 °C and then cooled to
45 °C for the crystallization. The resulting slurry was
reheated up to 65 °C to facilitate stirrability and to obtain
better filterability of crystals, and then it was cooled to 0-5
°C.
Acknowledgment
We thank Dr. Zanka of Chemical Development Labora-
tories for helpful discussions and documentation.
Received for review May 28, 2003.
OP0340661
706
•
Vol. 7, No. 5, 2003 / Organic Process Research & Development