Development of a practical and reliable synthesis of laquinimod

Article abstract of DOI:10.1021/op700001c

Laquinimod(5-chloro-1,2-dihydro-N-ethyl-4-hydroxy-1-methyl-2-oxo-N-phenyl- 3-quinoline carboxamide) is a drug candidate for treatment of Multiple Sclerosis. A short and industrially feasible process for the preparation of laquinimod starting from 2-amino-6-chlorobenzoic acid, in essentially four steps, is discussed. The key step is a novel reaction in which a methyl ester is converted to an amide in very high yield and with excellent purity. The present article elucidates the scale-up process along with safety aspects and the impurity profiles of the intermediates and product. Initial laboratory conditions are described as well as the changes made on transfer to pilot-plant scale.

Full text of DOI:10.1021/op700001c

Organic Process Research & Development 2007, 11, 674−680
Full Papers
Development of a Practical and Reliable Synthesis of Laquinimod
Johan Wennerberg,*,† Anders Bjo¨rk,‡ Tomas Fristedt,‡ Bo Granquist,† Karl Jansson,‡ and Ingela Τhuvesson‡
DuPont Chemoswed, R&D, P.O. Box 839, SE-201 80 Malmo¨, Sweden, and ActiVe Biotech Research AB,
P.O. Box 724, SE-220 07 Lund, Sweden
Scheme 1. Medicinal chemistry route to laquinimod
Abstract:
Laquinimod (5-chloro-1,2-dihydro-N-ethyl-4-hydroxy-1-methyl-
2-oxo-N-phenyl-3-quinoline carboxamide) is a drug candidate
for treatment of Multiple Sclerosis. A short and industrially
feasible process for the preparation of laquinimod starting from
2-amino-6-chlorobenzoic acid, in essentially four steps, is dis-
cussed. The key step is a novel reaction in which a methyl ester
is converted to an amide in very high yield and with excellent
purity. The present article elucidates the scale-up process along
with safety aspects and the impurity profiles of the intermedi-
ates and product. Initial laboratory conditions are described
as well as the changes made on transfer to pilot-plant scale.
Introduction
to laquinimod was essential to meet the production require-
ments. In this article we report a synthesis suitable for the
manufacturing of 1 at large scale with a novel conversion
of a methyl ester to an amide as the key step.
Multiple Sclerosis (MS) is an inflammatory disease of
the central nervous system (CNS) characterised by focal
leukocyte inflammation and demyelination resulting in nerve
cell dysfunction. At present â-Interferon (IFN-â) and glati-
ramer acetate are the major drugs used for treatment of MS.
A major drawback with â-Interferon is that about 20% of
the patients develop antibodies against the drug, and therefore
the search for new drugs are highly desired.
Laquinimod1 (1) has proven to significantly and dose-
dependently inhibit disease development in Experimental
Autoimmune Encephalomyelitis (EAE) which is an anti-
inflammatory autoimmune disease of the CNS and an
important experimental animal model for the study of MS.
Laquinimod also inhibits infiltration of CD4+ T cells into
the CNS in EAE induced animals. In contrast to IFN-â and
glatiramer acetate, laquinimod is administered orally and at
very low dose levels.2 During preclinical work and toxicology
testing, the need of laquinimod was supplied via a laboratory-
scale synthesis. Laquinimod will soon undergo phase III
studies, and process development of a scalable synthetic route
Results and Discussion
The first amounts of laquinimod used in preclinical and
toxicology studies were synthesised at laboratory scale via
a short three-step procedure (Scheme 1). Commercially
available 2-amino-6-chlorobenzoic acid was converted to
5-chloroisatoic anhydride (2) with phosgene in 1,4-dioxane
which after workup and drying gave 2 as a light brown
powder in almost quantitative yield. Anhydride 2 was then
converted to the desired compound 1 in a “one pot” two-
step procedure where 2 was alkylated with NaH/MeI in
dimethylacetamide followed by NaH-mediated reaction with
methyl 3-(N-ethyl-N-phenylamino)-3-oxopropionate (3) to
give 1 in 61% yield after workup and purification.
Although this synthesis seems elegant and straightforward,
it has several disadvantages when applied on a larger scale.
Phosgene is very poisonous and should be avoided unless
special reactors and equipment are available. Also the use
of sodium hydride is hazardous and should be avoided.
Beyond safety aspects, the present synthesis will require
synthesis of methyl 3-(N-ethyl-N-phenylamino)-3-oxopro-
pionate (3) from the expensive methyl malonyl chloride
which in turn will add one more step. In addition, this step
requires evaporation to dryness. The workup procedure and
purification in the final step are not very convenient including
washings with several different solvents. Small deviations
* Corresponding author. Telephone: +46 40 383316. Fax: +46 40 186395.
E-mail: johan.wennerberg@swe.dupont.com.
† DuPont Chemoswed.
‡ Active Biotech Research AB.
(1) (a) Noseworthy, J. H.; Wolinsky, J. S.; Lublin, F. D. Neurology 2000, 54,
1726. (b) Jo¨nsson, S.; Andersson, G.; Fex, T.; Fristedt, T.; Hedlund, G.;
Jansson, K.; Abramo, L.; Fritzon, I.; Pekarski, O.; Runstro¨m, A.; Sandin,
H.; Thuvesson, I.; Bjo¨rk, A. J. Med. Chem. 2004, 47, 2075. (c) Bjo¨rk, A.;
Jo¨nsson, S.; Fex, T.; Hedlund, G. U.S. Patent 6,077,851, 2000.
(2) (a) Polman, C.; Barkhof, F.; Sandberg-Wolheim, M.; Linde, A.; Nordle,
O¨ .; Nederman, T. Neurology 2005, 64, 987. (b) Jansson, K. U.S. Patent
6,875,869 B2, 2005.
674
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development
10.1021/op700001c CCC: $37.00 © 2007 American Chemical Society
Published on Web 06/30/2007

Scheme 2. Large-scale route to laquinimod
Figure 1. Intermediates in the first step and byproduct formed
in toluene.
2-amino-6-chlorotoluene could appear in the starting material.
However, this impurity, which probably originates from the
manufacture of 2-amino-6-chlorobenzoic acid, or its reac-
tion products were never seen in the product, probably due
to its good solubility in the mother liquor. Scale-up from
laboratory to pilot plant was straightforward for this step.
However, from laboratory trials it was noticed that efficient
cooling of the reflux condenser was important; otherwise
loss of acetyl chloride lowered the yield due to incomplete
conversion of 2a. Other solvents such as toluene or THF
gave lower yields or resulted in longer reaction times. Due
to the poor salt-solvating power of toluene the conversion
from hydrochloride 2b to carbamate 2a was extremely slow
which resulted in formation of the mixed anhydride 2c
(Figure 1). 2c seems unable to undergo cyclisation, and its
formation lowers the yield of to 2. Use of triethylamine to
circumvent formation of 2b led instead to extensive forma-
tion of 2c. In light of these findings 1,4-dioxane seemed to
be a superior solvent in this reaction despite its toxicological
properties.
The next step was to alkylate the amide. At laboratory
scale the reaction was performed with sodium hydride and
iodomethane in dimethylacetamide. The yield and purity
were both satisfying, but since sodium hydride is very
hazardous, especially at large scale and in connection with
solvents like DMA and DMF,4 an alternative reagent was
requested. Among tested bases potassium carbonate seemed
most promising. Laboratory trials demonstrated that 5 was
formed in good yield; apparently the N-H proton is acidic
enough to be abstracted by potassium carbonate. When
performed at large scale using iodomethane and potassium
carbonate in DMF at room temperature, the reaction was
complete after 72 h according to HPLC (Scheme 2). For
workup the mixture was diluted with 1 M hydrochloric acid.
Under these conditions the inorganic residues went into
solution and the product precipitated. The product was
isolated by centrifugation in 90% yield. This method was
safe and convenient; however, since the reaction is hetero-
geneous reaction time and stirring speed were of importance.
When the alkylation was performed at laboratory scale, the
reaction was completed within 12 h. In this case a magnetic
stirrer was used and the mixture was stirred quite vigorously
(approximately 800 rpm). The stirring speed in the pilot plant
reactor was much lower (approximately 50 rpm), and
consequently a prolonged reaction time was required. Also
in this case the assay was high, about 99%, and only two
impurities at low levels (0.07 and 0.14%) were found.
in the workup procedure will sometimes result in an oil,
instead of crystals. The purity of laquinimod was 94-95%
before and 97-98% after purification, methylester 4 (Scheme
2) being the most dominant impurity.
With this background a synthetic sequence more suitable
for large scale was desirable. We have previously synthesised
a similar compound, roquinimex,3 at multi-kilogram scale,
and parts of the present work are based on those experiences.
Ethyl chloroformate can serve as a carbonyl equivalent
instead of phosgene, and therefore the synthetic sequence,
aimed at large scale synthesis, can be started as shown in
Scheme 2. 2-Amino-6-chlorobenzoic acid with ethylchloro-
formate in 1.4-dioxane was heated at reflux temperature for
1 h after which acetyl chloride was added, followed by
heating for another 6 h (Scheme 2). Small-scale experi-
ments have revealed that approximately 60% of the car-
bamate intermediate 2a (Figure 1) was formed after a few
minutes according to NMR. Unreacted 2-amino-6-chloro-
benzoic acid was present in a less reactive and less sol-
uble hydrochloride form 2b (Figure 1), but this intermediate
was also transformed into the carbamate intermediate 2a
within 30 min and the mixture became clear. Addition of
acetyl chloride converted the carboxylic acid to an acid
chloride, which in turn underwent ring closure to the desired
substance, which precipitated after formation and was
isolated in 89% yield. The solubility of the product is very
low in 1,4-dioxane making the isolation via centrifugation
very convenient. The white product was very pure, more
than 99%, and except for 0.2% free chlorides no impur-
ities larger than 0.1% were found. NMR analysis of the
2-amino-6-chlorobenzoic acid indicated that up to 3%
(3) (a) Eriksoo, E.; Sandberg, E. B.; Stålhandske, L. J. T. U.S. Patent
4,547,511, 1985. (b) Sjovall, S.; Hansen, L.; Granquist, B. Org. Process
Res. DeV. 2004, 8, 802.
(4) Buckley, J.; Webb, R. L.; Laird, T.; Ward, R. J. Chem. Eng. News 1982,
60 (28), 5. De Wall, G. Chem. Eng. News 1982, 60 (37), 5.
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
675

Heterogeneous reactions may behave differently on different
scales.5 Granular potassium carbonate was used, and as a
consequence the reaction takes a longer time at a larger scale.
However, this reaction has been run in 1 g, 10 g, 100 g, 10
kg, and 30 kg, and no significant differences have been
observed except for the reaction times.
Figure 2. Possible impurity in the final product.
To form methyl ester 4, compound 5 was added to the
preformed enolate of dimethyl malonate at 80 °C in DMF
followed by stirring overnight at room temperature (Scheme
2). The enolate was initially formed with sodium hydride,
but this base was successfully replaced by sodium methoxide
at pilot-plant scale. Methanol is formed during the formation
of sodium dimethyl malonate and was removed together with
some DMF upon distillation of 10% of the solvent amount
before addition of the anhydride. If not removed, methanol
may react with 5-chloro-N-methylisatoic anhydride to give
the corresponding methyl anthranilate. The sodium dimethyl
malonate attacks 5-chloro-N-methylisatoic anhydride in a
regioselective manner. Carbon dioxide was eliminated, and
the intemediate underwent cyclisation to the desired product.
At the end of the reaction the product exists in the sodium
salt form, which precipitates in the acidic form after addition
of water and hydrochloric acid. Compound 4 was isolated
by centrifugation, the yield was 91%, and the purity was
almost 99%. It is also possible to carry out step 2 and 3 in
one pot. This was performed at laboratory scale by alkylation
of 2 with iodomethane and sodium hydride in dimethyl-
acetamide at 5-15 °C. To avoid 3-C and 4-O alkylation in
the next transformation, excess iodomethane was evaporated
under reduced pressure. Additional sodium hydride and
dimethyl malonate were added, while the mixture was heated
to 85 °C and kept at this temperature for 5 h. Acidic workup
and crystallisation from methanol afforded 4 in 71% yield.
No impurities were found. Although the one-pot procedure
seems convenient, it includes the use of hazardous sodium
hydride and also a crystallisation step. In light of these facts
the two-step procedure was the method of choice. As a matter
of fact the yield in the two-step procedure is more than 10%
higher (82%) than the yield in the corresponding one-pot
procedure (71%).
by slow addition of concentrated hydrochloric acid to acetic
anhydride under careful cooling (the reaction is very
exothermic). This mixture was diluted with glacial acetic
acid. With this method a HCl/acetic acid mixture with
approximately 1.5 M HCl and 1.4% H2O was obtained (HCl
in acetic acid is commercially available but quite expensive).
To this solution 4 was added, and the mixture was heated to
85 °C and kept at this temperature for 2 h (Scheme 2). The
mixture was cooled and diluted with methanol to facilitate
isolation. After centrifugation, washing, and drying, 7 was
obtained in 95% yield. The assay was 98.5% and decar-
boxylation product 6 could not be detected. However, 0.1%
of ester 4 was found. The present method, which releases
chloromethane7 instead of methanol as in usual hydrolysis
reactions, seems to be well suited for substrates which are
sensitive to decarboxylation.
With compound 7 in hand we were ready for the final
coupling step with N-ethylaniline. The classical method with
thionyl chloride and triethyl amine was tested at laboratory
scale and gave the desired compound 1 in 88% yield and
98% purity.2b However the impurity profile was not promis-
ing. Single impurities were well below the specification, but
numerous unknown impurities at low levels appeared as
indicated by HPLC. Manufacturing a pharmaceutical product
for clinical use with undefined byproducts is not very
convincing, and therefore other alternatives were searched.
Another common method for synthesising amides is the
DCC-mediated coupling of a carboxylic acid and an amine.8
A mixture of compound 7, N-ethylaniline, and DCC in
toluene was heated at 75-85 °C for 3-4 h (Scheme 2). After
cooling the product was isolated by centrifugation (filtration
at laboratory scale). The residue was washed with toluene
and dried. To remove N,N-dicyclohexylurea the dried product
was dissolved in aqueous sodium hydroxide. After adjust-
ment of pH to 6.5 N,N-dicyclohexylurea was filtered off.
Acidification of the filtrate gave a precipitate, which after
isolation and drying gave the desired compound in 62% yield.
The crude product was, however, contaminated with a
considerable amount of 6, but also 7 and 4 were found. To
solve the problem with these impurities a purification method
was developed. It was found that conversion to the crystalline
sodium salt using sodium hydroxide in ethanol could
significantly decrease the amounts of impurities and discol-
orations. This treatment yielded a product with sufficient
purity, and it is possible to even increase the purity with an
additional treatment. At pilot plant scale, 1 was stirred in
ethanolic sodium hydroxide for 2 h and the corresponding
sodium salt was isolated by filtration in 90-95% yield. The
sodium salt was then dissolved in water, and 1 precipitated
The next transformation, in which ester 4 is converted to
the carboxylic acid 7, was more difficult to perform.
Compound 7 is sensitive and seems to be prone to decar-
boxylate, especially under basic conditions.6 Decarboxylation
was total when 4 was heated to 80 °C in 1 M NaOH. Some
observations indicate that decarboxylation under acidic
conditions is favoured by high water content in the reaction
mixture. When hydrolysis was carried out in a mixture of
acetic acid and aqueous HCl, a large amount of decarboxy-
lated product 6 (Figure 2) was formed. To effectuate the
ester-acid transformation a rather unusual method was used.2b
A HCl/acetic acid mixture with low water content was made
(5) For examples with K2CO3, see: Conlon, D. A.; Drahus-Paone, A.; Ho, G.-
J.; Pipik, B.; Helmy, R.; McNamara, J. M.; Shi, Y.-J.; Williams, J. M.;
McDonald, D.; Descheˆnes, D.; Gallant, M.; Mastracchio, A.; Roy, B.;
Scheigetz, J. Org. Process Res. DeV. 2006, 10, 36. Mosely, J. D.; Bansal,
P.; Bowden, S. A.; Couch, A. E. M.; Hubacek, I.; Weinga¨rtner, G. Org.
Process Res. DeV. 2006, 10, 153.
(7) Evolved chloromethane was absorbed in methanolic potassium hydroxide.
(8) Bodanszky, M. Peptide Chemistry: A Practical Textbook; Springer: New
York, 1988.
(6) Jansson, K.; Fristedt, T.; Olsson, A.; Svensson, B.; Jo¨nsson, S. J. Org. Chem.
2006, 71, 1658.
676
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development

Table 1. Purification effect of salt conversiona
byproductsb
7
4
6
impure 1
one conversion
two conversions
5.0
0.95
0.26
0.3
0.15
0.05
10.0
0.49
0.07
Figure 3. Intermediate in the amide-forming step.
a Impure 1 was created from pure 1 by deliberately adding 7 and 6
to toluene in promoting the reaction. Although the developed
method seemed promising, there were still some drawbacks
when applied at a larger scale. The boiling point for trans-
decalin at 40-50 mbar is just above 100 °C, and the vacuum
has to be carefully controlled to avoid sudden vigorous
boiling, so that the mixture is not ejected from the reactor.
A system under vacuum is sensitive to leaks; particularly
this reaction is sensitive to moisture. To simplify synthesis
at large scale we decided to search for a solvent with a lower
boiling point in which the solubility of the product was still
very low. Small-scale experiments revealed that alkanes such
as n-heptane and n-octane were successful in promoting the
conversion of 4 into the desired compound 1.6 Distillation
with these solvents at atmosperic pressure gave similar results
as compared to trans-decalin, however, with much shorter
reaction times. When the reaction was performed in n-octane
at lab scale, the reaction was completed within 2 h. A period
of 6 h was required when the reaction was performed in
n-heptane. The production at pilot-plant scale was performed
with a 1:2 mixture of n-heptane and n-octane, and about one-
third of the solvent was distilled off during 5 h (Scheme 2).
This final method gave 1 in 96% yield, and as with the DCC-
method compound 1 was converted to its sodium salt as
described above. The assay in this case was as high as 99.9%.
The present method is superior in all aspects. Yield and purity
are excellent, and the method is simple and reliable when
applied at a large scale. More aspects regarding the mech-
anism of this reaction, which is believed to occur via the
ketene intermediate 8 (Figure 3), are reported elsewhere.6
The purity of the N-ethylaniline used is important. If other
N-alkylanilines or aniline itself are present as impurities,
these will react to give compounds similar to 1 which are
difficult to remove from the product. Our novel method for
amide formation has been used several times to manufacture
laquinimod at multi-kilogram scale as well as other sub-
stances in this class. The overall yield is 70% as compared
with 43% for the route using the DCC-coupling procedure.
corresponding to a level of 5% and 10%, respectively. bNumbers are in %.
upon addition of HCl. To demonstrate the effectiveness of
the salt-acid conversion, 1 was contaminated with 5 and
10% of 7 and 6, respectively. This impure material was then
purified Via conversion to the sodium salt. As seen in Table
1 the purifying effect is good especially for 6. This route,
using DCC and sodium salt conversion, was used in our first
manufacturing campaign of laquinimod at pilot-plant scale,
but we were still not satisfied, especially with the yield in
the DCC-coupling reaction.
Compounds similar to 1 have been synthesised at gram
scale by heating a mixture of a quinoline carboxylic ester
and an N-alkylated amine in toluene.2b,6 Toluene and the
formed alcohol are distilled off during the process, and the
distillation rate greatly influences the reaction outcome.
Because of a rather high solubility in toluene, the product
often had to be isolated by extraction with aqueous sodium
hydroxide or, if possible, precipitated with n-heptane. The
isolated products were often contaminated with a few percent
remaining ester. It occurred to us that if toluene was replaced
by another solvent that had a higher boiling point, thus
permitting a higher temperature and a faster reaction rate,
and in which the product had a very low solubility, the yield
might be increased. Our first choice was to use trans-decalin
(cis,trans-decahydronaphthalene, bp 189-191 °C). By heat-
ing a mixture of methylester 4 and N-ethylaniline in trans-
decalin at 100 °C and 40-50 mbar, an equilibrium between
4 and 1 is established. Methanol is formed during the reaction
between 4 and N-ethylaniline and is removed from the
reaction mixture by distillation. It is important to remove
the formed methanol in order to get a high yield of 1. This
equilibrium is well shifted towards the starting materials; in
fact, if 1 is heated with 1 equiv of methanol in toluene in a
sealed vessel, an almost complete conversion into 4 and
N-ethylaniline results. Similarly, if 1 equiv of water is
allowed to react with 1 in a similar fashion, 7 and decar-
boxylated compound 6 are the products. It is important that
no water, e.g., moist air, is allowed to enter the reaction
mixture, since that inevitably will result in formation of 6
and 7. It is also important that the reaction is not stopped
until almost all starting material has reacted. trans-Decalin
is promoting the reaction by allowing a high temperature at
a low pressure, thus facilitating efficient and complete
removal of methanol from the reaction vessel without
removing extensive amounts of trans-decalin. trans-Decalin
also seems to promote the reaction by providing a slightly
better solvation for 4 than for 1. The latter precipitates during
the reaction. The very low solubility of 1 in trans-decalin
may also explain the high purity of the product; as soon as
it is formed it precipitates and escapes decomposition
reactions. trans-Decalin is thus a superior medium compared
Conclusions
Overall, we have developed a safe, reliable, high-yielding,
and robust process for the synthesis of laquinimod. The
development includes a novel method for conversion of esters
into amides. The current route avoids undesired chemicals
like phosgene, sodium hydride, and dichloromethane. All
starting materials and reagents are cheap and easily available.
Isolation of intermediates and product is very convenient,
and the purity and impurity profiles of products and
intermediates are all very satisfying.
Experimental Section
General. HPLC analyses were performed using a Sym-
metry shield RP 8 (3.9 mm × 150 mm) from Waters
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
677

throughout. As mobile phase was used a gradient consisting
of A: 80% TFA, 0.1 M-20% CH3CN and B: 20% TFA,
0.1 M-80% CH3CN except for substance 1 where an isokratic
mixture of MeOH 40%, CH3CN 10%, and TFA 0.1 M 50%
was used. NMR spectra were recorded at 500.13 MHz
(proton) and 125.76 MHz (carbon), respectively. Assays were
determined by elemental analysis of the chlorine content in
each substance. Melting points are uncorrected. Thin layer
chromatography was performed on Merck precoated TLC
plates, and the acquisitions were visualised with UV light
or sprayed with a solution of p-methoxybenzaldehyde (26
mL), glacial acetic acid (11 mL), concentrated sulfuric acid
(35 mL), and 95% ethanol (960 mL). Solvents and reagents
were obtained from commercial sources and were used as
such without any further purification. All reactors used are
standard multipurpose equipment, either glass-lined or stain-
less steel. All reactions at pilot-plant scale are for safety
reasons routinely carried out under an atmosphere of
nitrogen.
5-Chloroisatoic Anhydride (2). In a glass-lined reactor,
equipped with a stirrer, reflux condenser, and gas absorption,
charged with 1.4-dioxane (83 kg) were added, under stirring,
2-amino-6-chlorobenzoic acid (10 kg, 58.3 mol) and ethyl
chloroformate (28 kg, 258 mol) under an atmosphere of
nitrogen. The mixture was heated at reflux temperature for
1 h after which it was cooled to 40 °C. Acetyl chloride (28
kg, 357 mol) was added at such a rate that the temperature
did not exceed 50 °C. Stirring was continued at 45 °C for
10 h whereafter the mixture was cooled below 10 °C. The
mixture was centrifuged, and the residue was washed with
toluene (2 × 10 L). Drying of the residue under reduced
pressure (8 mbar) at 50 °C to constant weight gave 10.2 kg
(89%) of the title compound as a white solid (assay 99.3%):
mp 272.5-272.9 °C; IR (KBr) 3450 (broad), 1780; 1H NMR
(DMSO-d6) δ 11.85 (s, broad, 1 H), 7.65 (t, 1 H, J ) 8.3
Hz), 7.31 (d, 1 H, J ) 8.3 Hz), 7.09 (d, 1 H, J ) 8.3 Hz);
13C NMR (DMSO-d5) δ 156.5, 146.7, 144.0, 136.5. 135.0,
125.7, 114.6, 108.1, 66.4. Anal. Calcd for C8H4ClNO3: C,
48.63; H, 2.04; N, 7.09. Found: C, 48.80; H, 2.23; N, 7.35.
5-Chloro-N-methylisatoic Anhydride (5). In a stainless
steel reactor, equipped with a stirrer, charged with DMF (100
L) were added, under stirring, 2 (9.5 kg, 48 mol), potassium
carbonate (7.2 kg, 52 mol), and iodomethane (8.7 kg, 61
mol). The mixture was stirred for 72 h at rt after which HCl
(200 L, 1 M) was added carefully (CAUTION! gas eVolu-
tion). The mixture was stirred for an additional 30 min and
was thereafter centrifuged. The residue was washed with
water (2 × 10 L) and dried to constant weight at 35 °C under
reduced pressure (6 mbar) which gave the title compound
(8.7 kg, 90%) as a white solid (assay 98.8%): mp 217.9-
218.2 °C; IR (KBr) 1780, 1725; 1H NMR (DMSO-d6) δ 7.78
(t 1 H, J ) 8.6 Hz), 7.42 (m, 2 H), 3.47 (s, 3 H); 13C NMR
(DMSO-d6) δ 155.5, 147.4, 144.6, 136.6, 135.4, 126.0, 114.1,
109.4, 32.4. Anal. Calcd for C9H6ClNO3: C, 51.09; H, 2.86;
N, 6.62. Found: C, 51.10; H, 2.83; N, 6.64.
charged with DMF (65 L) were added dimethyl malonate
(9.1 kg, 69 mol) and sodium methoxide (3.3 kg, 61 mol)
under an atmosphere of nitrogen. The mixture was heated
at 85 °C for 1.5 h after which approximately 10 L of solvent
were distilled off under reduced pressure (50 mbar). The
temperature was adjusted to 80 °C, and 5 (8.7 kg, 41 mol)
was added carefully during 10 min (CAUTION! foaming,
especially at the end of the addition). Another 20 L of solvent
were distilled off under reduced pressure (50 mbar). After
cooling to 20 °C the mixture was stirred overnight at ambient
temperature. Water (116 L) was added, and the mixture was
stirred until the solution became clear. HCl (37%, 10.5 kg)
was added slowly during 30 min until the pH was set to 1.
The mixture was cooled to 10 °C and stirred for 1 h at this
temperature to complete the precipitation. After centrifuga-
tion, the residue was washed with water (2 × 10 L) and
dried to constant weight at 50 °C at reduced pressure (8
mbar) which gave the title compound (10.0 kg, 91%) as a
white solid (assay 98.9%): mp 159.1-160.1 °C; IR (KBr)
3440 (broad), 1685; 1H NMR (CDCl3) δ 14.92 (s, 1 H), 7.52
(t, 1 H, J ) 8.5 Hz), 7.27 (m, 2 H), 4.04 (s, 3 H), 3.66 (s,
3 H); 13C NMR (CDCl3) δ 173.4, 172.7, 158.8, 143.6, 134.6,
133.4. 126.1, 113.3, 112.5, 98.1, 53.2, 30.1. Anal. Calcd for
C12H10ClNO4: C, 53.85; H, 3.77; N, 5.23. Found: C, 53.70;
H, 3.78; N, 5.40.
5-Chloro-1,2-dihydro-4-hydroxy-1-methyl-2-oxo-3-quin-
olinecarboxylic Acid (7). A glass-lined reactor, equipped
with a stirrer, reflux condenser, and gas absorption, charged
with acetic anhydride (35 kg, 343 mol) was cooled to 0 °C.
HCl (11.6 kg, 118 mol, 37%) was added slowly so that the
temperature was kept below 20 °C (CAUTION! The reaction
is strongly exothermic). Acetic acid (27.3 kg, 455 mol) and
4 (9.1 kg, 34 mol) were added, and the mixture was heated
at 85 °C for 2 h. The mixture was cooled to 10 °C, and
while it cooled, methanol (46 kg) was added at 65 °C. The
mixture was centrifuged, and the residue was washed with
methanol (2 × 12 kg). Drying to constant weight at 35 °C
under reduced pressure (7 mbar) gave the title compound
(8.2 kg, 95%) as a white solid (assay 98.5%): mp 294.9-
1
295.1 °C; IR (KBr) 3430 (broad), 1705; H NMR (CDCl3)
δ 15.78 (s, 1 H), 15.77 (s, 1 H), 7.65, (t, 1 H, J ) 8.1 Hz),
7.45(d, 1 H, J ) 8.1 Hz), 7.42 (d, 1 H, J ) 8.1 Hz), 3.79 (s,
3 H); 13C NMR (CDCl3) δ 174.0, 172.9, 164.3, 142.1, 135.4,
134.1, 127.7, 114.0, 113.8, 95.3, 30.5. Anal. Calcd for
C11H8ClNO4: C, 52.09; H, 3.18; N, 5.52. Found: C, 52.00;
H, 3.18; N, 5.66.
5-Chloro-1,2-dihydro-N-ethyl-4-hydroxy-1-methyl-2-
oxo-N-phenyl-3-quinoline Carboxamide (1), DCC-Medi-
ated from 7. To a glass-lined reactor charged with toluene
(35 kg) were added 7 (6.0 kg, 23.7 mol), N-ethylaniline (3,0
kg, 24.8 mol), and 1.3-dicyclohexylcarbodiimide (5.2 kg,
25.2 mol) under stirring. The mixture was heated to 80 °C
and kept at this temperature for 4 h under stirring. The
suspension was cooled to 10-15 °C and centrifuged. The
residue was washed with toluene (2 × 10 L) and dried to
constant weight at 35 °C under reduced pressure (8 mbar).
The dried substance was mixed with water (150 L), and
aqueous 2 M NaOH was added under stirring until the pH
5-Chloro-1,2-dihydro-4-hydroxy-1-methyl-2-oxo-3-quin-
olinecarboxylic Acid Methyl Ester (4). In a stainless steel
reactor, equipped with a stirrer and a distillation setup,
678
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development

was between 11 and 12. The clear solution was stirred for
30 min whereafter the pH was adjusted to 6.5 with aqueous
1 M HCl followed by additional stirring for 15 min. The
insoluble dicyclohexylurea was filtered off via a 3 µ filter.
Aqueous 5 M HCl was added to the filtrate over 10 min so
that the pH became ∼1. The mixture was stirred for 15 min
and was then left without stirring. After 2 h the mixture was
centrifuged, and the residue was washed with water until it
was free of chloride (free chloride was tested with 0.1 M
AgNO3 and 3 washes were usually sufficient). After drying
to constant weight (35 °C, 8 mbar) the product was obtained
as a white solid (5.2 kg, 62%): mp 201 (dec) °C; IR (KBr)
3450 (broad), 1640; 1H NMR (CDCl3) δ 12.25 (s very broad,
1 H), 7.39 (t, 1 H, J ) 8.2 Hz), 7.27-7.16 (m, 7 H), 3.99 (s
2 H), 3.27 (s, 3 H), 1.21 (s, 3 H); 13C NMR (CDCl3) δ 168.8,
165.7, 157.9, 142.7, 142.1, 132.8, 131.7, 128.5, 126.9, 126.7,
125.4, 113.3, 112.8, 105.1, 45.7, 29.8, 12.9; MS (ESI): m/z
222 (100). Anal. Calcd for C19H17ClN2O3: C, 63.96; H, 4.80;
N, 7.85. Found: C, 63.70; H, 4.75; N, 8.11.
dimethyl malonate (38.2 g, 0.29 mol) in one portion. The
mixture was heated to 85 °C (gas evolution started at 40
°C) and was kept at that temperature for 5 h. The mixture
was cooled and water (4 L) was added. The pH was adjusted
to 1.5-2 with hydrochloric acid (1 M), and the mixture was
stirred for 1 h. The mixture was filtered, and the filter cake
was washed with water (200 mL) and n-heptane (100 mL)
and dried at 50 °C under reduced pressure (5 mbar).
Crystallisation from methanol gave the title compound as
white crystals (48 g, 71%). No impurities were found
according to NMR analysis.
5-Chloroisatoic Anhydride (2) with the Phosgene
Method. To a stirred and cooled slurry of 2-amino-6-
chlorobenzoic acid (100 g, 0,57 mol) in 1,4-dioxane (500
mL) was added, dropwise, phosgene (62.5 g, 0.63 mol)
(CAUTION! Very toxic) in 1,4-dioxane (500 mL) at 12-15
°C. The mixture was stirred at 20 °C, and then after it cooled
to 5 °C, sodium acetate trihydrate was added in one portion
followed by a rapid addition of ice-cold water (3 L). The
mixture was stirred for 2 h. The product was filtered off
and washed with water (2 L), MeOH/H2O 4:6 (1 L), and
n-heptane (300 mL). Drying over P2O5 at reduced pressure
(5 mbar) gave the title compound (112 g, 99%) as a light-
brown powder.
Methyl 3-(N-Ethyl-N-phenylamino)-3-oxopropionate
(3). To a stirred solution of N-ethylaniline (98 g, 0.81 mol)
in acetone (900 mL) was added triethylamine (127 mL, 0.91
mol) under an atmosphere of nitrogen. The mixture was
cooled to 2 °C, and methyl malonyl chloride (100 mL, 0.91
mol) was added at a rate so that the temperature did not
exceed 20 °C. The mixture was stirred at rt for 2 h and
thereafter concentrated under reduced pressure. The residue
was dissolved in toluene (700 mL) and washed with water
(700 mL), 0.5 M H2SO4 (700 mL), and saturated aqueous
NaHCO3 (3 × 700 mL). The solution was dried over MgSO4
and concentrated under reduced pressure, and the residue
was coevaporated twice with CHCl3. Drying under reduced
pressure gave the title compound (169 g, 94%) as a light
brown oil; IR (film) 1745, 1700; 1H NMR (CDCl3) δ 7.47-
7.37 (m, 3 H), 7.21 (d, 2 H, J ) 7.6 Hz), 3.56 (q, 2 H, J )
10.4 Hz), 3.53, (s, 3 H), 3.17 (s, 2 H), 1.14 (t, 3 H, J ) 10.4
Hz); 13C NMR (CDCl3) δ 168.2, 165.4, 141.7, 129.9, 128.4,
128.4, 52.3, 44.3, 41.7, 12.9. Anal. Calcd for C12H15NO3:
C, 65.14; H, 6.83; N, 6.33. Found: C, 64.4; H, 6.82; N, 6.19.
5-Chloro-1,2-dihydro-N-ethyl-4-hydroxy-1-methyl-2-
oxo-N-phenyl-3-quinoline Carboxamide (1) in One Pot
from 2. NaH (19.1 g, 0.56 mol. 70% in mineral oil) was
washed twice with n-pentane and added in portions, under
stirring, to a solution of 2 (101 g, 0.51 mol) in dimethyl-
acetamide (1 L) at 4 °C under an atmosphere of nitrogen at
such a rate that the temperature did not exceed 20 °C. After
addition of NaH, the temperature was adjusted to 6 °C and
iodomethane (38.2 mL, 0.61 mol) was added dropwise during
10 min. The cooling bath was removed, and the reaction
was left stirring at rt overnight. To remove excess iodo-
methane the mixture was kept at reduced pressure (10 mbar)
for 2 h. NMR analysis indicated that the alkylation was
5-Chloro-1,2-dihydro-N-ethyl-4-hydroxy-1-methyl-2-
oxo-N-phenyl-3-quinoline Carboxamide (1) SOCl2-medi-
ated from 7 is described in reference 2b.
5-Chloro-1,2-dihydro-N-ethyl-4-hydroxy-1-methyl-2-
oxo-N-phenyl-3-quinoline Carboxamide (1), directly from
Methylester 4. To a glass-lined reactor charged with
n-heptane (60 L) and n-octane (120 L) were added N-
ethylaniline (6.8 kg, 56.1 mol) and 5 (10.0 kg, 37.4 mol).
The mixture was heated to reflux (112 °C) under a slow
stream of nitrogen. During 5 h approximately 60 L of solvent
were distilled off (temperature rises to 122 °C). The mixture
was kept at 120 °C for 3 h after which it was cooled to 25
°C. n-Heptane (40 L) was added, and the mixture was stirred
overnight at ambient temperature. After centrifugation and
washing with n-heptane (2 × 15 L), the substance was dried
to constant weight at 50 °C and reduced pressure (8 mbar).
The yield of the title compound was 12.8 kg (96%), and the
purity was 98.5 area%.
Conversion of 1 to Sodium Salt. To a stainless steel
reactor charged with ethanol (90 L) was added 1 (12.8 kg,
35.9 mol), and the mixture was stirred for 15 min. To the
mixture was added 10 M NaOH (4.0 L, 40 mol), and stirring
was continued for 1.5 h. The product was centrifuged,
washed with ethanol (2 × 12 L), and dried to constant weight
at 35 °C under reduced pressure (7 mbar) to give the product
as white crystals (12.95 kg, 95%). Assay was 99.9%.
5-Chloro-1,2-dihydro-4-hydroxy-1-methyl-2-oxo-3-quin-
olinecarboxylic Acid Methyl Ester (4) in One Pot from
2. Compound 2 (50 g, 0.25 mol) was dissolved in dimethyl-
acetamide (500 mL), and the solution was cooled to 5 °C
under an atmosphere of nitrogen. Sodium hydride (9.6 g,
0.28 mol, 70% in mineral oil) was added, under stirring, so
that the temperature was kept below 15 °C. Iodomethane
(43.5 g, 0.31 mol) was added dropwise, the cooling bath
was removed, and the mixture was stirred at ambient
temperature for 2 h. Excess iodomethane was removed by
keeping the mixture under reduced pressure (15 mbar) for 1
h. Additional sodium hydride (9.6 g, 0.28 mol, 70% in
mineral oil) was added in portions followed by addition of
Vol. 11, No. 4, 2007 / Organic Process Research & Development
•
679

complete. NaH (19.1 g, 0.56 mol. 70% in mineral oil) was
washed twice with n-pentane and added in portions, under
stirring, followed by addition of malonamide 3 (126 g, 0.56
mol). The mixture was heated to 85 °C during 2 h (gas
evolution occurred at 35-40 °C), and the mixture was kept
at this temperature for another 5 h. The mixture was cooled
to rt and left overnight. The mixture was kept at reduced
pressure (10 mbar) for 1 h after which it was cooled to 5
°C. Methanol (1 L) was added, and the temperature was
adjusted to 10 °C. The cooling bath was removed, and HCl
(1340 mL, 1 M) was added quickly under stirring. H2O (7
L) was added, forming an emulsion which crystallised
slowly. After 2 h of crystallisation the mixture was stirred
vigorously for 1 h at rt and another 1 h at 0 °C. The product
was isolated by filtration; washed with H2O (7 L), MeOH/
H2O (1:1, 5 L), n-heptane (2 L), and n-pentane (1 L); and
dried over P2O5 at reduced pressure. Yield 134 g (74%) and
assay 94.3%.
The compound was purified as follows: The impure
product was dissolved in 1 M NaOH (850 mL) and filtered
through Celite and washed with CH2Cl2 (3 × 200 mL). The
pH was adjusted to 6.5 with 2 M HCl, and the brown and
turbid solution was filtered through Celite followed by
acidification to pH 1.5 and stirring for 1.5 h. The precipitated
product was centrifuged and washed with water (5L) and
dried over P2O5 at reduced pressure (5 mbar). 110 g (61%)
of product was obtained with an assay of 97.6%
Physical data for compound 6 can be found in ref 6.
Received for review January 3, 2007.
OP700001C
680
•
Vol. 11, No. 4, 2007 / Organic Process Research & Development