Process development and pilot-scale synthesis of cefotetan

Article abstract of DOI:10.1021/op049914m

Strategies that were adopted during the process development of Cefotetan in order to achieve a cost-effective commercial-scale synthesis are described herein. These included replacement of the trifluoroacetic acid used for cleavage of the benzhydryl ester and the development of an alternative synthetic route. This work led to improvement of both the impurity profile and the yield of the process. The pilot-scale synthesis of Cefotetan is described in detail in the Experimental Section. The scaled-up process has been successfully used for the commercial manufacture of Cefotetan since 1983.

Full text of DOI:10.1021/op049914m

Organic Process Research & Development 2004, 8, 915−919
Process Development and Pilot-Scale Synthesis of Cefotetan
Masaharu Fujimoto, Tetsuya Maeda, Keisuke Okumura, Mitsuru Uda, Makoto Nakamura, Teruya Kashiwagi, and
Takashi Tsunoda*
Chemical Technology Labs, Yamanouchi Pharmaceutical Co., Ltd., 160-2, Akahama, Takahagi-shi,
Ibaraki, 318-0001, Japan
Abstract:
Scheme 1. Synthesis of Cefotetan (10)
Strategies that were adopted during the process development
of Cefotetan in order to achieve a cost-effective commercial-
scale synthesis are described herein. These included replacement
of the trifluoroacetic acid used for cleavage of the benzhydryl
ester and the development of an alternative synthetic route. This
work led to improvement of both the impurity profile and the
yield of the process. The pilot-scale synthesis of Cefotetan is
described in detail in the Experimental Section. The scaled-up
process has been successfully used for the commercial manu-
facture of Cefotetan since 1983.
Introduction
The overall process development of Cefotetan synthesis
1
was outlined previously. This paper describes the modifica-
tions that were necessary to achieve a practical pilot-plant
synthesis that could be scaled up for the commercial
2
manufacture of this substance. Cefotetan (YM09330) is a
semisynthetic cephamycin antibiotic that was discovered and
developed by the Yamanouchi Pharmaceutical Co., Ltd.,
Japan, and has been on the market since 1983. This
compound shows strong wide-spectrum antibacterial activity
and has high stability against â-lactamase. This active
pharmaceutical ingredient has been manufactured on an
industrial scale at the Takahagi plant in Ibaraki prefecture,
Japan, for more than 20 years. The synthetic route to
Cefotetan is shown in Scheme 1. The starting material,
Oganomycin GG (1), is itself the product of a biochemical
to protect the carboxylic acid functionality in cephalosporin
chemistry, was relatively expensive; therefore, a cheaper
alternative to this reagent was sought. Third, the potential
for developing an alternative synthetic route, with the
exception of the rearrangement reaction, was explored.
3
process and is converted into Cefotetan (10) after a total of
eight reactions.
4
In general, the complex syntheses and high dosages that
are required for production of an antibiotic lead to high
manufacturing costs, which represent a significant hurdle that
must be overcome in order to bring these products to the
market. Cefotetan was no exception; however, three target
areas were identified in which cost improvement could be
achieved. First, as the initial pilot-plant scale-up only afforded
a 60% yield for the conversion of 1 to 6 and 40% for the
conversion of 6 to 10, methods for improving these yields
were investigated. Second, trifluoroacetic acid, which has
been traditionally used to cleave the benzhydryl ester used
Results and Discussion
Development of the Process from 1 to 6. The esterifi-
2 2
cation of 1 by diphenyldiazomethane in CH Cl produced 2
quantitatively, with nitrogen as a byproduct of the process.
The reaction mixture could be used directly in the next step.
Difficulties arose at this stage because the three intermediates,
3
, 4, and 5, were highly unstable with moisture or at room
(4) An alternative synthetic route, which involves the coupling of 5 and 2-(tert-
butoxycarbonyl carbamoylmethylene)-1,3-dithietane-4-carboxylic acid (11)
was developed. 10, produced by the deprotection of the coupling product
and the overall yield of this alternative synthetic route, which did not involve
any rearrangement reaction, was around 40% from 1.
*
Author for correspondence. E-mail: tsunoda.takashi@yamanouchi.co.jp.
Telephone: +81 293 23 4485. Fax: +81 293 24 2708.
(1) Fujimoto, M. J. Synth. Org. Chem. Jpn. 1997, 57, 374.
(2) Iwanami, M.; Maeda, T.; Fujimoto, M.; Nagano, Y.; Nagano, N.; Yamazaki,
A.; Shibanuma, T.; Tamazawa, K.; Yano, K. Chem. Pharm. Bull. Jpn. 1980,
2
8, 2629.
(3) Osono, T.; Watanabe, S.; Saito, T.; Gushima, H.; Murakami, K.; Susaki,
K.; Takamura, S.; Miyamoto, T.; Oka, Y. J. Antibiot. 1980, 33, 1074.
1
0.1021/op049914m CCC: $27.50 © 2004 American Chemical Society
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
915
Published on Web 11/03/2004

temperature. Compound 6 was comparatively stable and was
isolated with a 60% yield from 1. HPLC revealed that the
three major byproducts obtained in the CH Cl extract of 6
2 2
were approximately 10% of the 7-epimer (6a) (Scheme 2),
2
approximately 5% of the ∆ -isomer (6b) (Scheme 3), and 2
or 3% of the intermediate (2).
Scheme 2. Epimerization of the 7-MACA ester (5)
Figure 2. Comparison of the isomerization rate of GG ester
2) under different conditions.
(
measured. These data showed that a weak base, such as Py,
was more effective at preventing isomerization and also
revealed that the temperature was crucial. The stability of 5
2 2
in the CH Cl solution into which it was extracted was
Having predicted the mechanism by which these byprod-
ucts were produced, the rates of epimerization of 5 (Scheme
) after the methanolysis reaction of 4 were measured at three
compared at three different temperatures, to determine the
most suitable temperature for the acylation reaction. The
results, which are shown in Figure 3, indicated that 5 was
highly sensitive to temperature and that the acylation reaction
should be kept at below -20 °C.
5
2
different temperatures (-10, -24, and -40 °C). Since 5 was
easily epimerized in methanol at room temperature, the rates
were monitored using the following method. First, after the
methanolysis reaction at -45 °C for 4 h, seven samples of
the reaction mixture were maintained for 0, 1, and 3 h at
each temperature. Next, this allowed conversion to 6, which
was then analyzed using HPLC. The results, which are shown
in Figure 1, revealed that methanolysis should be conducted
at below -40 °C in order to significantly reduce the rate of
epimerization.
Scheme 3. Isomerization of the GG ester (2)
Figure 3. Comparison of the stability of the 7-MACA ester
2 2
(5) in CH Cl solution at different temperatures.
It was necessary to optimize the methods used to monitor
the PCl reaction, as the intermediate 3 reacted readily with
water to give 2. Since the reaction was complete within 1 to
h, its progress was therefore monitored by measuring the
amount of residual 2 using TLC within 10 min. The
procedure involved adding 1 mL of the reaction mixture of
3 to methanol, to give 5 via the intermediate 4. Hydrochloric
5
2
2
The rate of ∆ -isomerization (Scheme 3) was subsequently
investigated at three different temperatures in the presence
of pyridine (Py) and at one temperature with triethylamine
(
TEA) as the base. The results are shown in Figure 2. The
required amount of amine was added to a CH Cl solution
of 2, and the rate of formation of the ∆ -isomer was
f
acid was then added to decompose 5, as its R value was
similar to that of 2, and the organic layer was spotted onto
small TLC plates.
2
2
2
Experiments were then performed to determine the most
suitable conditions for these processes with respect to
temperature, reaction time, and the reagents used. This
resulted in an optimization of the overall yield for converting
1
to 6 to 90%. In addition, the production of each of the
three major byproducts 6a, 6b, and 2 was reduced to <1%.
Cleavage of the Protected Carboxylic Acid 6. Although
several alternative reagents such as chloromethylmethyl ether,
chloromethylbenzyl ether, diphenylbromomethane, and tert-
butyldimethyl silyl chloride were investigated, none com-
pared well in yield with the performance of diphenyldi-
azomethane for protection of the carboxylic acid. The
Figure 1. Comparison of the epimerization rate of the
-MACA ester (5) at different temperatures.
2
7
difficulty with the esterification was due to the ∆ -isomer-
916
•
Vol. 8, No. 6, 2004 / Organic Process Research & Development

ization of 2 under basic conditions. Regarding the cleavage
of the benzhydryl ester, trifluoroacetic acid, which has been
generally chosen for the cleavage, was considered to be
unsuitable for the commercial production because of its high
information, as shown in Figure 5. The pH was lowered using
a stepwise process from 8.5 to 8.0 to 7.5, over a 45 h reaction
period. The final equilibrium of 9/10 was 5:95 at 1 °C.
price. A cheaper alternative was sought, and the phenomenon
6
that HCl gas dissolves in CH
2
Cl
2
at low temperatures was
exploited to convert 6 to 7 almost quantitatively at temper-
atures <-30 °C. On the other hand, the AlCl /nitromethane
3
7
cleavage of benzhydryl ester derivatives was published
around the same time. The performance of the two methods
was compared. The results indicated that the HCl/CH
2 2
Cl
method performed better than the AlCl /nitromethane method
3
on a large scale. A reliable supplier of diphenyldiazomethane
was able to be sourced, and so the protection and cleavage
issue was concluded.
Figure 5. Rearrangement of 9 at 1 °C.
Development of the process from 6 to Cefotetan (10).
After 6 was converted to 7 using the method described above,
In the early stages of the process development, the mixture
of 9 and 10 was extracted into ethyl acetate/2-butanol in the
ratio 3:1. Treatment of the resulting mixture with ethanol
afforded 10. However, the mixed solvent showed a low
efficiency for dissolving 10, and the purification using
ethanol was not particularly effective. Although absolute
methylethyl ketone also dissolved 10 inefficiently, hydrated
methylethyl ketone produced significantly better results.
Therefore methylethyl ketone was chosen as the extraction
solvent, and after concentration, methanol was added to
precipitate (10) as an amorphous solid. Purification with
methanol proved to be highly effective at removing impuri-
ties; however, the precipitate had the appearance of cotton
wool, and filtration by centrifuge was extremely slow. It was
therefore filtered using a filter press with a wide filtration
area. The wet cake from the filter press was then converted
into a suspension in ethanol, filtered again using the filter
press, and dried to give high purity 10. The overall yield of
7
3
was extracted into aqueous NaHCO solution and subse-
quently reacted with 8, still in aqueous solution, to give 9
quantitatively. In the final step, Cefotetan (10) was formed
from its isomer (9) via an interesting rearrangement reaction.
As shown in Figure 4, 9 remained mainly in the isothiazole
form at pH 10 but was mostly rearranged in the dithietane
form (10) at pH 9. Although the time required to reach
equilibrium was shorter at higher pH and increased temper-
atures, the degradation of 9 and 10 was also accelerated under
these conditions.
1
0 from 6 was about 68%.
As the filtrate contained approximately 22% of 10 and
% of 9, a second recovery process was implemented.
5
Although 10 could be recovered by concentrating the filtrate,
9 could not. Further investigation revealed that the Ca salts
of both 9 and 10 could be recovered from the filtrate by the
Figure 4. Equilibrium between 9 and 10 in 0.1 M phosphate
buffer at 37 °C.
addition of triethylamine and CaCl
comprised a mixture of both 9 and 10 in an approximately
:8 ratio. Adding the recovered Ca salt to the next batch of
the rearrangement-reaction mixture at a pH of 8.0 increased
the yield of 10 from 6 to 87%.
2
. The recovered Ca salt
The equilibrium shifted towards 10 at pH 7.0 to 7.5, and
2
-
the presence of an anion of a weak acid, such as HCO
accelerated the rearrangement of 9 to 10. Therefore, the
rearrangement of 9 was carried out in aqueous NaHCO , with
3
,
2
3
the pH adjusted several times using an acid. When HCl or
phosphoric acid was used in the pH adjustment, both 9 and
Conclusions
1
0 partially precipitated from the solution and the pH could
not be accurately measured until the precipitate had redis-
solved. As an alternative, the introduction of CO was found
This research program revealed that Cefotetan could be
manufactured on a commercial scale using the route outlined
in Scheme 1, with an overall yield for the conversion of 1
to 10 of 78%. Development was therefore scaled-up, and
through collaboration with engineering specialists, a dedi-
cated automated plant was completed in early 1983. Manu-
facture has subsequently continued without the need for
further changes to the methods described.
2
to be useful in the adjustment of pH and for acceleration of
the rearrangement, because it did not produce any precipitate.
Furthermore, both 9 and 10 were comparatively stable around
pH 7 at 1 °C. The most suitable conditions for the
rearrangement reaction were determined on the basis of this
(5) Lunn, W. H. W.; Burchfield, R. W.; Elzey, T. K.; Mason, E. V. Tetrahedron
Lett. 1974, 14, 1307.
Experimental Section
Materials and Instrumentation. Oganomycin GG was
manufactured in the Takahagi plant of the Yamanouchi
(
(
6) Ahmed, W.; Gerrard, W.; Maladkar, V. K. J. Appl. Chem. 1970, 20, 109.
7) Tsuji, T.; Kataoka, T.; Yoshioka, M.; Sendo, Y.; Nishitani, Y.; Hirai, S.;
Maeda, T. Tetrahedron Lett. 1979, 30, 2793.
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
917

Pharmaceutical Co. Ltd., Japan. DDM was obtained from
the Koei Chemical Co., Ltd., Japan. Tripotassium 4-carboxy-
and the reaction mixture was stirred for 3 h at -35 to -38
°C to produce 7. The completion of the reaction was
confirmed using TLC. A mixed solution of water (330 L)
and n-butanol (220 L) was added to the reaction mixture at
-38 °C, and the temperature of the reaction mixture rose to
3-hydroxy-5-mercapto isothiazole was obtained from the
Tateyama Kasei Co., Ltd., Japan. All other chemicals were
1
obtained from the usual commercial suppliers. H NMR
spectra were recorded on a JEOL PX-100 spectrometer, with
the chemical shifts given in ppm relative to TMS at δ ) 0.
HPLC was carried out using a Waters model 440 liquid
chromatograph. A Shimadzu model UV-300 was used as UV
spectrophotometer.
-4 °C. After the mixture was stirred, the CH
separated and the aqueous layer was back-extracted with
CH Cl (142 L). A cold solution of water (621 L) and
NaHCO (28.1 kg) was added to the combined CH Cl layers,
2 2
Cl layer was
2
2
3
2
2
and the mixture was stirred at 0 °C in order to extract 7 into
Preparation of Diphenylmethyl 7â-Bromoacetamido-
the aqueous phase. The aqueous layer was washed twice with
7
r-methoxy-3-[(1-methyl-1H-tetrazol-5-yl)thiomethyl]-3-
CH
2
Cl
2
(248 L, 110 L). The pH of the aqueous phase was
, and 32 kg of NaHCO were
cephem-4-carboxylate (6) from Oganomycin GG (1).
then adjusted to 6.7 using CO
2
3
Oganomycin GG (1) (80.0 kg; 169 mol) and a 50% solution
added to the aqueous solution. An aqueous solution of
tripotassium 4-carboxy-3-hydroxy-5-mercaptoisothiazole (8)
(44.5 kg, 153 mol) in water (115 L) was cooled to 6 °C and
added dropwise to the aqueous solution of 7 at 0 °C. As the
aqueous solution of 8 was strongly basic, the pH of the
of DDM (72.3 kg, 372 mol) in CH
Cl (744 L) at <-5 °C. The reaction mixture was stirred
for 2 d at 2 °C to produce 2. Pyridine (18.38 kg, 232 mol)
and PCl (44.1 kg, 211 mol) were added to the reaction
2 2 2
Cl were added to CH -
2
5
mixture at 2 °C and warmed to 20 °C. The reaction mixture
was stirred for 2 h at 18 to 21 °C to produce 3 and then
cooled to -60 °C. Methanol, which was cooled to -45 °C
2
mixture was adjusted using CO to between pH 7.0 and 8.6
during the addition. The reaction of 7 and 8 was monitored
using HPLC, and the addition of 8 was continued until the
complete disappearance of 7 to produce 9. The pH of the
(
208 L), was added to the reaction mixture at <-45 °C to
produce 4. The reaction mixture was stirred for 4 h at -45
C to produce 5. Next, 25% ammonia water (76.3 kg, 1122
reaction mixture was initially adjusted with CO
and stirring was continued for 3 h at 0 °C. The pH was then
adjusted again, using CO , to a value of 8.0. The recovered
2
to pH 8.5,
°
mol) was added to water (216 L) and cooled to 2 °C. The
cooled ammonia-water solution was added to the reaction
mixture at -45 °C, and the mixture was stirred for 40 min
2
Ca salt from the previous batch (39 kg) was added to the
reaction mixture. This was initially stirred at pH 8.0 to 7.8
for 22 h at 0 °C and then at pH 7.5 for 24 h at 0 °C. Cold
methylethyl ketone (1,260 L) and cold 20% HCl (218 kg)
were added to the reaction mixture, and the product was
extracted into the methylethyl ketone. The organic layer was
separated, and the aqueous layer was back-extracted with
methylethyl ketone (290 L). The moisture content of the
combined organic layers was adjusted to 8.3% with water
addition, activated charcoal (6 kg) was added to the organic
layer, and the slurry was stirred for 30 min at 15 °C. The
activated charcoal was then removed by filtration. The
moisture of the filtrate was adjusted to 9.5% with further
water addition. The organic solution was concentrated to
about 250 L under vacuum at <15 °C. MeOH (1220 L) was
added to the residue, and the slurry was stirred for 12 h at
temperatures between 0 and -5 °C. The precipitate was
filtered using a filter press at -5 °C, and the wet cake from
the filter press was converted into a slurry in EtOH (850 L).
The product was filtered again at -5 °C, and the wet cake
was dried in a vacuum at 10 °C to yield 76 kg (132 mol) of
the desired product (10). The purity of 10 was 99.42% as
at approximately 0 °C. The CH
washed twice with cold water (280 L × 2) and cooled to
60 °C. Pyridine (22.1 kg, 279 mol) and bromoacetyl
bromide (54.8 kg, 271 mol) were added to the CH Cl layer
2 2
Cl layer was separated and
-
2
2
at -60 °C, and the temperature of the reaction mixture
increased to approximately -40 °C over a 10 min period.
The reaction mixture was stirred for 1 h at <-20 °C, then
cold 5.43% H
mixture was stirred for 1 h at approximately 0 °C. The CH
Cl layer was separated and washed with water (288 L). The
CH Cl was almost completely removed by concentration
2 4
SO (297 kg, 165 mol) was added and the
2
-
2
2
2
under vacuum at a temperature <15 °C. Benzene (512 L)
was added to the residue in order to promote the crystal-
lization of 6. After the slurry of 6 was stirred for 15 h at 10
°
C, the crystals were collected by centrifugal filtration and
washed with benzene. The crystals of 6 contained a benzene
molecule from the solvent of crystallization. Quantitative UV
analysis showed the wet cake of 6 dissolved in CH
2 2
Cl (275
L) to be equivalent to 110 kg (156 mol) of dry crystals. The
1
overall yield was 90%. H NMR (100 MHz, CDCl
s, 5H, -OCH , -CH -), 3.78 (s, 3H, NCH ), 3.88 (s, 2H,
BrCH CO), 4.35 (q, 2H, -CH S-), 4.04 (s, 1H), 6.92 (s,
H, -OCH<), 7.36 (m, 17H, diphenyl, NH, C ).
3
) δ 3.57
2
(
3
2
3
HPLC area. The major impurities were 0.27% of ∆ -isomer
1
2
2
of 10 and 0.20% of 9. The total yield was 87% from 6. H
1
6
H
6
NMR (100 MHz, DMSO-d
(q, 2H, -CH -), 3.92 (s, 3H, NCH
), 5.12 (s, 1H), 5.16 (s, 1H, S >CH-), 7-8 (b, 2H, -NH
6
) δ 3.42 (s, 3H, -OCH
3
), 3.62
S-
),
Preparation of 7â-[[4-(Carbamoyl carboxymethylene)-
,3-dithietan-2-yl]carboxamido]-7r-methoxy-3-[[(1-meth-
2
3
), 4.29 (q, 2H, -CH
2
1
2
2
yl-1H-tetrazol-5-yl)thio]methyl]-3-cephem-4-carboxylic Acid
9.63 (s, 1H, -NH-). HPLC was performed with UV
detection at a wavelength of 254 nm using an ODS 150 mm
× 4 mm column and eluting with 0.1 M phosporic acid/
acetonitrile/acetic acid/methanol, 17/1/1/1. TLC was per-
formed with silica gel F254 plates (Merck), eluting with ethyl
acetate/n-hexane 2/1.
(
10) from 6. A mixed solution of CH Cl (770 L) and
2
2
n-butanol (8.8 L) was cooled to -60 °C, and HCl gas (55.5
kg, 1520 mol) was introduced at a temperature <-50 °C.
2 2
The solution of 6 in CH Cl was added dropwise into the
HCl/CH Cl solution at temperatures from -60 to -35 °C,
2
2
918
•
Vol. 8, No. 6, 2004 / Organic Process Research & Development

Recovery of the Ca Salt. A solution of CaCl
2
(0.67 kg)
Acknowledgment
in water (0.67 L) and triethylamine (1.19 kg) was added to
the first MeOH filtrate at 5 °C, and the mixture was stirred
for 30 min at 0 °C. The resulting precipitate was removed
using a filter press. The second EtOH filtrate was concen-
trated to 168 L under vacuum at around 20 °C. The residue
and second MeOH filtrate were combined, and a solution of
We are grateful to Dr. Toshiyasu Mase, Dr. Shuichi
Sakamoto (Yamanouchi Pharmaceutical Co., Ltd.), and Dr.
Donal Murphy (Yamanouchi Ireland Co., Ltd.) for their
helpful advice. We are also grateful to the Analysis Labo-
ratories (Yamanouchi Pharmaceutical Co., Ltd., Tsukuba,
Japan) for spectral measurements.
2
CaCl (7.4 kg) in water (7.4 L) and triethylamine (13.43 kg)
was added to the combined filtrates. The mixture was stirred
for 15 h at 0 °C, and the resulting precipitate was collected
using a filter press. The wet cake was dried under vacuum
at 10 °C to yield 40 kg of the Ca salt.
Received for review April 27, 2004.
OP049914M
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
919