Synthesis of [1-15N,2-13C]-labeled difloxacin

Article abstract of DOI:10.1007/s00706-009-0158-y

The synthesis of [1-¹⁵N,2-¹³C]-difloxacin, an arylfluoroquinolone antibacterial agent, is reported. As a crucial initial step, the starting materials ethyl 2,4,5-trifluorobenzoylacetate, [formyl- ¹³C]-triethyl orthoformate, and [¹⁵N]-4-fluoroaniline were reacted to ethyl [¹⁵N,3-¹³C]-3-(4-fluoroanilino)-2-(2,4, 5-trifluorobenzoyl)acrylate. After cyclization and ester cleavage, the resulting intermediate was reacted with 1-methylpiperazine to [1-¹⁵N,2- ¹³C]-1-(4-fluorophenyl)-6-fluoro-7-(4-methyl-1-piperazinyl)-1, 4-dihydro-4-oxoquinoline-3-carboxylate, i.e., [1-¹⁵N,2- ¹³C]-difloxacin. The overall yield was 62% based on the non-labeled and 43% based on the labeled starting materials (both used in 1.4 molar excess). The product was identified by ¹H-, ¹³C-, and ¹⁵N-NMR spectroscopy and by cochromatography (TLC, HPLC) with an authentic reference; its purity (HPLC) was above 98%. Prior to synthesis of [1-¹⁵N,2-¹³C]-difloxacin, non-labeled difloxacin was synthesized in order to optimize procedures and to identify and characterize all intermediates.

Full text of DOI:10.1007/s00706-009-0158-y

Monatsh Chem (2009) 140:1221–1227
DOI 10.1007/s00706-009-0158-y
ORIGINAL PAPER
Synthesis of [1-15N,2-13C]-labeled difloxacin
Burkhard Schmidt Æ Ramona Schumacher-Buffel Æ
¨
Brigitte Thiede Æ Andreas Schaffer
Received: 25 September 2008 / Accepted: 21 January 2009 / Published online: 29 May 2009
Ó Springer-Verlag 2009
Abstract The synthesis of [1-15N,2-13C]-difloxacin, an
arylfluoroquinolone antibacterial agent, is reported. As a
crucial initial step, the starting materials ethyl 2,4,5-tri-
fluorobenzoylacetate, [formyl-13C]-triethyl orthoformate,
and [15N]-4-fluoroaniline were reacted to ethyl [15N,3-
13C]-3-(4-fluoroanilino)-2-(2,4,5-trifluorobenzoyl)acrylate.
After cyclization and ester cleavage, the resulting inter-
mediate was reacted with 1-methylpiperazine to
[1-15N,2-13C]-1-(4-fluorophenyl)-6-fluoro-7-(4-methyl-1-
piperazinyl)-1,4-dihydro-4-oxoquinoline-3-carboxylate,
i.e., [1-15N,2-13C]-difloxacin. The overall yield was 62%
based on the non-labeled and 43% based on the labeled
starting materials (both used in 1.4 molar excess). The
(for the chemical structure, see Scheme 1) is effective
against a wide range of gram-positive and gram-negative
aerobes and anaerobes byacting as a DNA gyrase inhibitor. It
has been developed and approved for veterinary use in pigs,
cattle, sheeps, goats, poultry, and dogs [1–3]. In contrast to
pesticides, which are used in agriculture by spraying con-
siderably large areas, most veterinary pharmaceuticals are
applied to individual organisms. Contamination of the
environment during medication itself is consequently neg-
ligible. Nonetheless, compounds can be introduced into the
environment by three main mechanisms. One route is via
sewage, sewage treatment plants, and surface waters, which
applies, e.g., for unused pharmaceuticals that are sometimes
disposed of in the sewage system. Direct input into soils may
occur via excretion of feces and urine by grazing animals.
The main route, however, is via manure, which is widely
used as an agricultural fertilizer. Thus, portions of antibiotics
remaining non-metabolized or metabolites exhibiting
residual biological activity are excreted by animals, end up in
manure, and may subsequently contaminate agricultural
soils [4–8].
1
product was identified by H-, 13C-, and 15N-NMR spec-
troscopy and by cochromatography (TLC, HPLC) with an
authentic reference; its purity (HPLC) was above 98%.
Prior to synthesis of [1-15N,2-13C]-difloxacin, non-labeled
difloxacin was synthesized in order to optimize proce-
dures and to identify and characterize all intermediates.
Keywords Isotopic labeling ꢀ Antibiotics ꢀ
NMR spectroscopy ꢀ Difloxacin ꢀ Arylfluoroquinolone
Until recently, little attention has been paid to the intro-
duction of veterinary pharmaceuticals into the environment
[8]. However, veterinary antibiotics have been detected in
soils as the unintentional result of using manure [5–8]. There
are only a few reports on groundwater contamination,
although antibiotics have been detected in surface waters due
to runoff from fields fertilized with manure [6, 9]. According
to these findings, it is considered reasonable by regulatory
authorities (legislation) to study both the effects and fate of
veterinary antobiotics, such as difloxacin, in agricultural
soils fertilized with manure from treated livestock [6, 8, 10,
11]. Until now, data on the latter issue have been scarce [5].
The present synthesis of 13C- and 15N-labeled difloxacin
was performed in the context of a larger research unit
Introduction
Fluoroquinolones have found wide application as antibac-
terial agents in human and veterinary medicine. Difloxacin
B. Schmidt (&) ꢀ R. Schumacher-Buffel ꢀ B. Thiede ꢀ
¨
A. Schaffer
Institute of Biology V (Environmental Research),
RWTH Aachen University, Worringerweg 1,
52056 Aachen, Germany
e-mail: burkhard.schmidt@post.rwth-aachen.de
123

1222
B. Schmidt et al.
Scheme 1
funded by the German Research Foundation, which deals
with the effects and fate of the veterinary antibiotics sulf-
adiazin and difloxacin in soil after their introduction via
manure of correspondingly medicated pigs. Results con-
cerning the fate of sulfadiazin have already been published
[12, 13]. Utilization of 13C-labeled sulfadiazin as an ana-
lytical reference was crucial for these experiments [13]. A
similar reference labeled with stable isotopes was thus
required for the study with difloxacin. In soil, sulfadiazin
[12] and difloxacin [14] demonstrated a considerable ten-
dency to rapidly form so-called non-extractable residues.
These are thought to consist of covalent and non-covalent
associations of the antibiotic or its metabolites with soil
organic and inorganic matter that are not disintegrated by
simple extraction procedures. In order to obtain informa-
tion on the nature of these associations regarding their
possible environmental impact, solid or liquid phase 13C
NMR spectroscopy using a suitable 13C-labeled tracer was
considered promising [15]. Thus, the projected difloxacin
labeled with stable isotopes was also intended for field of
application.
Results and discussion
Synthesis planning
According to the literature [16–20], a key step in the
synthesis of arylfluoroquinolone antibacterial agents is
treatment of an appropriately substituted (methyl or) ethyl
benzoylacetate (1, in case of difloxacin) with triethyl
orthoformate (2a) to give the one-carbon homologue
enol ether. The resulting intermediate is allowed to react
with aniline 3a to yield the enamine keto ester 4a, which
123

Synthesis of [1-15N,2-13C]-labeled difloxacin
1223
upon cyclization gives ethyl 1,4-dihydro-4-oxoquinoline-3-
carboxylate 5a. This product is then hydrolyzed to the
carboxylic acid 6a and reacted with an amine (1-methylpi-
perazine, 7, in case of difloxacin) yielding arylfluoroquinolone
(Scheme 1).
was obtained in oily form. Removal of acetic anhydride
was essential in order to avoid the side reaction to N-acetyl
4-fluoroaniline during the subsequent synthetic step. Under
argon cover gas, the oily intermediate was dissolved in
dichloromethane and converted to ethyl 3-(4-fluoroanili-
no)-2-(2,4,5-trifluorobenzoyl)acrylate (4a) by reaction with
4-fluoroaniline (3a; molar excess: 1.4 regarding 1) [16, 17].
Crucial intermediate 4a was purified by silica gel column
chromatography, and characterized and identified by thin-
As outlined before, preparation of difloxacin labeled
with stable isotopes was performed in order to obtain both
an analytical reference (for mass spectrometry-based
analysis, i.e., LC-MS) and a suitable tracer for 13C NMR
studies on the nature of the so-called non-extractable res-
idues emerging from the transformation of difloxacin in
soil. Multiple labeling, such as a uniformly labeled aro-
matic ring, was preferable in the case of the analytical
reference (with a distinctly different mass compared to the
non-labeled molecule), while labeling of positions with
high stability towards, e.g., microbial transformation was
required for the tracer compound. The initial steps of di-
floxacin metabolism in humans [21, 22] and dogs [23] were
shown to occur at the piperazine moiety, which subse-
quently can be removed completely. In contrast, the
remaining carbon skeleton appears to be comparatively
stable. Since a similar initial transformation pattern was
suspected in soil, labeling was regarded to be the most
favorable in the quinolone or 4-fluorophenyl moiety of
difloxacin. This conclusion is supported by data published
on the metabolism of the closely related fluoroquinolones
ciprofloxacin and enrofloxacin by the brown rot fungus
Gloeophyllum striatum [24, 25].
1
layer chromatography (TLC) [26], H and 19F NMR, and
GC-EIMS. Both NMR spectra demonstrated the presence
of (E)- and (Z)-isomers of 4a (with regard to the C2–C3
olefinic bond), the ratio of which varied among different
experiments. Prominent 1H NMR signals were those of the
NH and H-3 protons, respectively appearing as doublets at
d = 12.50/11.09 and 8.50/8.41 ppm. Peaks at d = 4.12/
4.09 (q) and 1.12/0.99 ppm (t) indicated the signals of the
ethyl group, whereas those of all aromatic protons emerged
as a multiplet with d = 7.50–6.80 ppm. The 1H NMR
spectrum of 4a correlated with published spectra of
similar compounds [17, 19, 20]. In addition to (E)- and
(Z)-isomers, the 19F NMR spectrum showed that four
fluorine atoms were present in the molecule. These
exhibited expected chemical shifts (main isomer given) at
d = -142.7 (q), -130.8 (q), -115.8 (q), and -115.3 ppm
(s), 19F–19F spin–spin coupling, and corresponding con-
stants. As examined by gas chromatography-mass spec-
troscopy with electron impact ionization (GC-EIMS), the
mass spectrum of 4a was similar to that of subsequent
intermediate 5a (including GC retention time) supposedly
due to cyclization in the injector (250 °C) of the chromato-
graphic system.
Ethyl 2,4,5-trifluorobenzoylacetate (1) and 4-fluoroani-
line (3) are not commercially available as (uniformly)
ring-13C-labeled derivatives. The same holds for 2,4,5-tri-
fluorobenzoic acid, which can be converted to 1 [16, 17,
20]. Uniformly 13C-labeled aniline is available; conversion
of this precursor to 3, however, was beyond the scope of
the present study. Thus, labeling of difloxacin was carried
out using [formyl-13C]-triethyl orthoformate (2b). In order
to increase the mass of the labeled compound, [15N]-4-
fluoroaniline (3b) was additionally utilized as an aniline
component. This ultimately resulted in [1-15,2-13C]-di-
floxacin, which was considered suitable as a tracer and
analytical reference.
Cyclization of 4a to ethyl 1-(4-fluorophenyl)-6,
7-difluoro-1,4-dihydro-4-oxoquinoline-3-carboxylate (5a)
was executed using NaH in dry tetrahydrofurane under argon
atmosphere in accordance with a published procedure [16,
17]. Intermediate product 5a was examined without further
purification by TLC [26] by 1H, 13C, and 19F NMR, as well as
by GC-EIMS, and proved to be sufficiently pure for the
subsequent synthetic reactions. The 1H and 19F NMR spectra
unequivocally demonstrated the cyclization of 4a to 5a. The
1
NH signal (of 4a) was absent in the H NMR spectrum of
Synthesis of non-labeled difloxacin and identification
of intermediates
intermediate 5a, while only signals of three flourine atoms
emerged in the 19F NMR spectrum of 5a (d = -138.3,
-126.9, and -108.5 ppm) as compared to 4a (four signals).
In the 1H NMR spectrum, aromatic protons H-5, H-8, H-20/
H-60, and H-30/H-50 appeared with high resolution, and
demonstrated expected H–1H and H–19F spin-spin cou-
pling patterns and constants. The most noticeable feature of
the 13C NMR spectrum of intermediate 5a was the doublets
of carbon atoms C4a, C8, C8a, C10, C20, C30, C40, C50, and
C60 and the double doublets of carbon atoms of C5, C6, and
C7, which were due to 13C–19F spin–spin coupling. Coupling
For the optimization of methods and unequivocal identifi-
cation of intermediates, non-labeled difloxacin was
prepared prior to labeling with 13C and 15N (Scheme 1). A
mixture of ethyl 2,4,5-trifluorobenzoylacetate (1), triethyl
orthoformate (2a; molar excess: 1.4), and acetic anhydride
was reacted as described in Refs. [16, 17]. After evapora-
tion of the byproduct (ethyl acetate), remaining starting
materials, 2a, and acetic anhydride, the resulting intermediate
1
1
123

1224
B. Schmidt et al.
constants were 13C–19F: 252–256 Hz, 13C–C–19F: 14–
23 Hz, and 13C–C–C–19F: 2–9 Hz. Unexpectedly, the 13C
signal of carbon atom C8 consisted only of a doublet—in
contrast to C5, which gave rise to a double doublet caused by
spin-spin coupling with both F-6 and F-7. This was probably
due to insufficient resolution of the spectrum with regard to
C8. At least, this assumption is supported by the observation
that the peaks of the doublet of C8 were comparatively broad.
Both the 1H and 13C NMR spectrum agreed with published
spectra of fluoroquinolones (other than difloxacin) [17, 19,
HPLC analysis, the purity of the compound was above
98%.
Synthesis of [1-15N, 2-13C]-labeled difloxacin
After optimization of methods, the synthesis of
[1-15N,2-13C]-difloxacin was executed as displayed in
Scheme 1. Procedures and data given before correspond
with maximum yields; these were used for the preparation
of the labeled derivative. Intermediates 4b, 5b, and 6b
were identified by TLC, and 4b and 5b additionally by GC-
EIMS. In order to minimize inevitable losses, only the
projected [1-15N,2-13C]-difloxacin was thoroughly identi-
?
20, 27, 28]. The EIMS of 5a exhibited a molecular ion Mꢀ
at m/z = 347 and a base peak at m/z = 275 derived from
?
Mꢀ –COOC2H4 fragmention. Fragment ion m/z = 219
1
was tentatively thought to result from fragmentation
?
fied by H, 13C, and 15N NMR spectroscopy and HPLC
Mꢀ –COOC2H4–CH2O–CN. Other fragment ions were con-
examination—in addition to TLC.
sistent with the structure of 5a.
Ethyl [15N, 3-13C]-3-(4-fluoroanilino)-2-(2,4,5-trif-
luorobenzoyl)acrylate (4b) was prepared from 1,
[formyl-13C]-triethyl orthoformate (2b), and [15N]-4-fluo-
roaniline (3b). Based on compound 1, the yield of 4b was
99% after column chromatography. Cyclization gave ethyl
[1-15N, 2-13C]-1-(4-fluorophenyl)-6,7-difluoro-1,4-dihy-
dro-4-oxoquinoline-3-carboxylate (5b; yield: 89.4% based
on 1). During TLC, intermediates 4b and 5b co-chro-
matographed with the corresponding non-labeled
derivatives (4a and 5a, respectively). Both 4b and 5b
showed the same MS and retention time with GC-EIMS
examination (see discussion on 4a and 5a before). Char-
Intermediate 5a was transformed to the free carboxylic
acid 6a by refluxing in a mixture of HCl, acetic acid, and
water according to a recently published procedure [18].
Hydrolysis under basic conditions [16] was less effective,
and isolation of the product was more complicated. 1-(4-
Fluorophenyl)-6,7-difluoro-1,4-dihydro-4-oxoquinoline-3-
1
carboxylate (6a) was identified by TLC [26], H, and 19F
NMR spectroscopy. With the exception of the ethyl group
(1H NMR), the spectra of ester 5a and free acid 6a were
rather similar, agreeing with published data on fluoro-
quinolones [17, 19, 20, 27, 28]. Finally, intermediate 6a
was reacted with 1-methylpiperazine (7) using triethyl-
amine as a base to give the projected non-labeled difloxacin
(8a) according to a published method [18]. The product
was identified (by comparison with an authetic reference)
?
acteristic fragments emerged at m/z = 277 (Mꢀ –
COOC2H4; base peak) and m/z = 219 (277-COOC2H4–
CH2O–13C15N) besides m/z = 349, which corresponded
with the molecular ion. The fact that both the non-labeled
compound 5a and labeled derivative 5b exhibited the same
fragment at m/z = 219 was regarded as proof for the sus-
pected fragmentation pattern. Cleavage of ester 5b gave
[1-15N, 2-13C]-1-(4-fluorophenyl)-6,7-difluoro-1,4-dihy-
dro-4-oxoquinoline-3-carboxylate (6b); the yield was 62%
based on ethyl benzoylacetate 1.
1
by TLC [26], HPLC, H, and 13C NMR. 19F NMR spec-
troscopy of 8a was not possible due to its poor solubility in
alternative solvents, and trifluoroacetic acid was required.
The 1H NMR of 8a resembled that of 6a. Main differences
were the signals of protons H-5 and H-8, which were both
doublets (spin–spin coupling with F-6) as compared to
double doublets in case of 6a (spin–spin coupling with F-6
and F-7). Additionally, the signals of the protons of the
piperazinyl moiety emerged in the range of d = 3.90–
Using triethylamine as catalyst, intermediate 6b was
reacted with 1-methylpiperazine (7) to the projected
[1-15N,2-13C]-difloxacin (8b). The yield of 8b amounted to
62% based on 1 introduced into the reaction sequence,
while its purity was above 98% as determined by HPLC
analysis. Compound 8b co-chromatographed (TLC, HPLC)
with the non-labeled 8a. Its 1H and 13C NMR spectra were
similar to the corresponding spectra of 8a; two exceptions
however, were observed, which were due to the 1-15N- and
2-13C-labeling of 8b. The signal of hydrogen atom H-2
emerged as a doublet at d = 9.37 ppm, resulting from
1H-13C spin-spin coupling (J = 190.8 Hz). Similarly, the
signal of carbon atom C2a appeared as a large doublet
(d = 152 ppm) in the 13C NMR spectrum 8b, which con-
trasted with both the size of remaining 13C signals in this
1
3.20 ppm; the H NMR pattern obtained was supported by
literature data on such ring systems [29]. A similar trend
was observed in the 13C NMR spectrum of 8a. The signals
of carbon atoms C4a, C5, C6, and C7 appeared as doublets,
which exhibited the expected 13C–19F spin-spin coupling
constants (in case of C6, 250 Hz, and smaller constants
with other carbon nuclei). In contrast to carbon C4a, a
(broad) singlet, and not a doublet, was observed as a signal
of carbon C8. A similar phenomenon was observed with
intermediate 5a (see discussion above). As expected, the
signals of the piperazinyl part of the molecule were
observed at d = 56.5, 48.9, and 46.0 ppm. According to
123

Synthesis of [1-15N,2-13C]-labeled difloxacin
1225
spectrum and the singlet signal of carbon atom C2 in the
spectrum of non-labeled 8a. Both effects were brought
about by 13C labeling of position C2 of the molecule and
by spin-spin coupling (J = 42 Hz) between 2-13C and
1-15N. Additionally, a 15N NMR spectrum of compound 8b
was recorded, which demonstrated a doublet at d =
-215.7 ppm and a 15N–13C spin–spin coupling constant of
J = 38.4 Hz.
were system I: n-hexane:diethyl ether 1:1 (v/v) and system
II: methanol:iso-propanol:NH3 (conc.) 2.5:2.5:1 (v/v/v)
[26]. Electron impact mass spectra (EIMS) were deter-
mined using a Hewlett–Packard 5890 Series II gas
chromatograph (Agilent Technologies, Waldbronn,
Germany) equipped with an FS-Supreme-5 column
(25 m 9 0.25 mm, 0.23 lm film thickness; CS, Langer-
wehe, Germany). The chromatograph was connected to a
Hewlett-Packard 5971 A MSD (mass selective detector)
operated in scan mode (m/z = 50–500) with 70 eV. Carrier
gas was He and injection volume 1 mm3. HPLC was
performed on a Beckman System Gold personal chro-
matograph (Munich, Germany) with Programmable
Solvent Module 126 and Diode Array Detector Modul 168.
Analyses were performed at 20 °C on a reversed phase
column (CC 250/4 Nucleodur Sphinx-RP 5 lm; Macherey-
Nagel, Du¨ren, Germany) Elution was carried out with
solvents A (0.05 M H3PO4 in H2O) and B (CH3CN): A/B
95:5 (v/v) for 5 min, linear 20 min gradient to A/B 50:50,
linear 10 min gradient to A/B 10:90, return to initial con-
ditions in 5 min, and isocratic A/B 95:5 for 5 min. HPLC
analyses were carried out with 100 mm3 injection volume
at a flow rate of 1 cm3 min-1 with detection of difloxacin
at 260, 286, and 300 nm. 1H, 13C, 19F, and 15N NMR
spectra were recorded using Bruker DPX300 and AV600
instruments; d in ppm up- and downfield of internal stan-
dard (CH3)4Si.
Conclusion
[1-15N,2-13C]-Difloxacin, i.e., [1-15N,2-13C]-1-(4-fluoro-
phenyl)-6-fluoro-7-(4-methyl-1-piperazinyl)-1,4-dihydro-4-
oxoquinoline-3-carboxylate, was synthesized from non-labeled
ethyl 2,4,5-trifluorobenzoylacetate, and the labeled starting
materials [formyl-13C]-triethyl orthoformate and [15N]-4-
fluoroaniline. Overall yield was 62% based on the non-
labeled and 42% based on the labeled starting materials (each
introduced using a 1.4 molar excess). The isotope-labeled
difloxacin is useful as an internal standard for mass spec-
trometry-based analytical procedures, which are required
for quantification of the arylfluoroquinolone antibacterial
agent in excreta of medicated livestock and in soils fertilized
with manure from treated animals. The labeled compound
can also be used as a tracer for 13C NMR studies on the
nature of non-extractable residues arising from difloxacin
in soil.
Synthesis of non-labeled-difloxacin and [1-15N,2-13C]-
labeled-difloxacin
Experimental
The procedures for the synthesis of non-labeled difloxacin
were similar to those used for the synthesis of
[1-15N,2-13C]-difloxacin (8b) (see below and Scheme 1).
All intermediates (4a, 5a, 6a) and the final product (8a)
were identified by NMR spectroscopy and occasionally by
supplementary methods as follows.
Materials
Ethyl 2,4,5-trifluorobenzoylacetate (1) was supplied by
ChemPur (Karlsruhe, Germany), difloxacin [1-(4-fluoro-
phenyl)-6-fluoro-7-(4-methyl-1-piperazinyl)-1,4-dihydro-
4-oxoquinoline-3-carboxylate, 8] as hydrochloride by
Fluka (Buchs, Switzerland; VetranalÒ, analytical standard).
Triethyl orthoformate (2) and 4-fluoroaniline (3) as well as
Ethyl 3-(4-fluoroanilino)-2-(2,4,5-trifluorobenzoyl)-
acrylate (4a, C18H13F4NO3)
1H NMR (300 MHz, CDCl3), mixture of (E)- and (Z)-
isomer (chemical shifts were not assigned to isomers):
d = 12.50/11.09 (d, 1H, NH, J = 13.5 Hz), 8.50/8.41 (d,
1H, H-3, J = 13.5 Hz), 7.50–6.80 (m, 6 H, H-20, H-30,
H-50, H-60, H-30, H-600), 4.12/4.09 (q, 2H, CH2CH3,
J = 7.0 Hz), 1.12/0.99 (t, 3H, CH2CH3, J = 7.0 Hz)
ppm; 19F NMR (282 MHz, CDCl3), mixture of (E) and
(Z) isomer (main isomer presented): d = -142.7 (dd, F-400,
J = 22 Hz, J = 16 Hz), -130.8 (dd, F-500, J = 22 Hz,
J = 6 Hz), -115. 8 (dd, F-200, J = 16 Hz, J = 6 Hz),
-115.3 (s, F-40) ppm; EIMS: Same spectrum as compound
5a (see below) due to cyclization of 4a to 5a in the injector
of the GC system.
¨
1-methylpiperazine were purchased from Riedel de-Haen
(Selze, Germany) and Fluka (Buchs, Switzerland),
respectively. [Formyl-13C]-triethyl orthoformate (2b; 13C
enrichment: 99.4 atom%, chemical purity: 98%) and [15N]-
4-fluoroaniline (3b; 15N enrichment: 95 atom%, chemical
purity: 98%) were provided by Campro Scientific GmbH
(Berlin, Germany).
Analytical methods
Thin-layer chromatography (TLC) was executed on silica
gel plates (SIL G-25 UV254, thickness: 0.25 mm; Mache-
rey-Nagel, Du¨ren, Germany). The solvent systems used
123

1226
B. Schmidt et al.
Ethyl 1-(4-fluorophenyl)-6,7-difluoro-1,4-dihydro-4-
(2b; 7.179 mmol), and 2.259 g acetic anhydride (22.13
mmol) was heated to 130 °C for 120 min. During the
reaction, the emerging ethyl acetate evaporated and was
separated by a condenser. After cooling to 22 °C, remaining
portions of acetic anhydride and 2b were removed under
high vacuum. The remaining oily orange product was put
under an argon atmosphere and dissolved in 25 cm3 of dry
dichloromethane. To this solution, 0.798 g [15N]-4-fluoro-
aniline (3b; 0.68 cm3, 7.179 mmol) was added, and the
resulting mixture was stirred for 60 min at 22 °C under
argon. Remaining portions of 3b and dichloromethane were
removed under high vacuum. The resulting brown solid
crude product was dissolved in chloroform and adsorbed
onto 10 g of silica gel 60 (0.063–0.2 mm; Fluka). This
adsorbate was laid on a 100-g column of silica gel 60
(3 9 34 cm, solvent: petrol ether:diethyl ether 95:5 v/v).
The crude product was fractionated using a gradient of petrol
ether:diethyl ether and fractions of 100 cm3 as follows: 95:5
(v/v) 10 fractions, 90:10 (v/v) 10 fractions, 80:20 (v/v) 10
fractions. Fractions 16–27 were combined and dried under
high vacuum. Yield of 4b was 1.780 g (99% based on 1); the
intermediate was examined by TLC (solvent system I:
Rf = 0.38) and GC-MS. EIMS: same spectrum as com-
pound 5b (see below) due to cyclization in the injector of the
GC system.
oxoquinoline-3-carboxylate (5a, C18H12F3NO3)
1H NMR (300 MHz, CDCl3): d = 8.40 (s, 1H, H-2), 8.17
(dd, 1H, H-5, J = 10.5 Hz, J = 8.7 Hz), 7.51 (dd, 2 H,
H-20, H-60, J = 8.5 Hz, J = 4.6 Hz), 7.34 (dd, 2H, H-30,
H-50, J = 8.5 Hz, J = 8.3 Hz), 6.72 (dd, 1H, H-8,
J = 11.0 Hz, J = 6.2 Hz), 4.31 (q, 2H, CH2CH3,
J = 7.1 Hz), 1.34 (t, 3H, CH2CH3, J = 7.1 Hz) ppm; 13C
NMR (75 MHz, CDCl3): d = 172.6 (C4), 165.0 (COOEt),
163.3 (d, C40, J = 252 Hz), 153.3 (dd, C6, J = 256 Hz,
J = 15 Hz), 149.0 (C2), 149.0 (dd, C7, J = 252 Hz,
J = 14 Hz), 137.7 (d, C8a, J = 9 Hz), 136.2 (d, C10,
J = 3 Hz), 129.5 (d, C20, C60, J = 9 Hz), 125.5 (d, C4a,
J = 5 Hz), 118.0 (d, C30, C50, J = 23 Hz), 115.2 (dd, C5,
J = 19 Hz, J = 2 Hz), 111.3 (C3), 106.5 (d, C8,
J = 23 Hz), 61.1 (OCH2CH3), 14.3 (OCH2CH3) ppm;
19F NMR (282 MHz, CDCl3): d = -138.3 (d, F-7,
J = 24 Hz), -126.9 (d, F-6, J = 24 Hz), -108.5 (s, F-40)
?
ppm; EIMS: m/z (relative intensity) = 347 (7, Mꢀ ), 318
?
?
?
(1, Mꢀ –C2H5), 302 (22, Mꢀ –OC2H5), 275 (100, Mꢀ –
COOC2H4), 258 (12, 275-OH), 247 (9, 275-CO), 245 (10,
275-CH2O), 226 (7, 245-F), 219 (18, 245-CN).
1-(4-Fluorophenyl)-6,7-difluoro-1,4-dihydro-4-
oxoquinoline-3-carboxylate (6a, C18H8F3NO3)
1H NMR (300 MHz, DMSO-d6): d = 8.78 (s, 1H, H-2),
8.34 (dd, 1H, H-5, J = 10.4 Hz, J = 8.5 Hz), 7.77 (dd,
2H, H-20, H-60, J = 8.7 Hz, J = 4.9 Hz), 7.54 (dd, 2 H, H-
30, H-50, J = 8.7 Hz, J = 8.7 Hz), 7.26 (dd, 1H, H-8,
J = 11.4 Hz, J = 6.6 Hz) ppm; 19F NMR (300 MHz,
DMSO-d6): d = -137.4 (d, F-7, J = 24 Hz), -125.6 (d,
F-6, J = 27 Hz), -110.3 (s, F-40) ppm.
Ethyl [1-15N,2-13C]-1-(4-fluorophenyl)-6,7-difluoro-
1,4-dihydro-4-oxoquinoline-3-carboxylate
13
17
(5b, C CH12F135NO3)
Intermediate 4b was dissolved in 40 cm3 of dry tetrahy-
drofurane under argon, and 234 mg of an oily suspension
(60%) of NaH (5.847 mmol) was added. The mixture was
heated to 50 °C for 240 min under argon and was then
cooled to 0 °C. After addition of 0.8 cm3 of acetic acid
(13.644 mmol), the mixture was concentrated to about
2 cm3 (rotary evaporator), when 100 cm3 of water was
added. The sedimented orange brown solid was separated
by filtration and dried (high vacuum). Yield of 5b was
1.535 g (89% based on 1). The product was examined by
TLC (solvent system I: Rf = 0.22) and GC-MS. EIMS: m/z
Difloxacin [1-(4-fluorophenyl)-6-fluoro-7-(4-methyl-1-
piperazinyl)-1,4-dihydro-4-oxoquinoline-3-carboxylate]
(8a, C21H19F2N3O3)
1H NMR (300 MHz, trifluoroacetic acid-d1): d = 9.11 (s,
1H, H-2), 8.22 (d, 1H, H-5, J = 12.2 Hz), 7.52 (m, 2H,
H-20, H-60), 7.36 (dd, 2H, H-30, H-50, J = 7.6 Hz,
J = 7.6 Hz), 6.74 (d, 1H, H-8, J = 6.4 Hz), 3.9–3.6 (m,
4H, piperazinyl 2- and 6-CH2), 3.5–3.2 (m, 4H, piperazinyl
3- and 5-CH2), 3.00 (s, 3H, H3C-N) ppm; 13C-NMR
(151 MHz, trifluoroacetic acid-d1): d = 174.0 (C4), 172.4
(COOH), 167.5 (d, C40, J = 255 Hz), 157.8 (d, C6,
J = 258 Hz), 152.2 (C2), 150.5 (d, C7, J = 10 Hz),
143.9 (C8a), 137.3 (C10), 130.9 (d, C20, C60, J = 7 Hz),
120.9 (d, C30, C50, J = 25 Hz), 118.1 (d, C4a, J = 12 Hz),
114.2 (d, C5, J = 25 Hz), 109.2 (C8), 106.2 (C3), 56.5
(C300, C500), 48.9 (C200, C600), 46.0 (N–CH3) ppm.
?
?
(relative intensity) = 349 (9, Mꢀ ), 320 (1, Mꢀ –C2H5),
?
?
304 (22, Mꢀ –OC2H5), 277 (100, Mꢀ –COOC2H4), 260
(12, 277-OH), 249 (10, 277-CO), 247 (12, 277-CH2O), 228
(9, 247-F), 219 (17, 247-13C15N).
[1-15N,2-13]-1-(4-Fluorophenyl)-6,7-difluoro-1,4-
dihydro-4-oxoquinoline-3-carboxylate
13
15
(6b, C CH8F135NO3)
Intermediate 5b was dissolved in 10.5 cm3 of a mixture of
water, HCl (conc.), and acetic acid (conc.) 1:1:1 (v/v/v),
and was refluxed for 180 min. With cooling to 22 °C, the
product sedimented, and was separated by filtration and
washed with ethanol. The light brown product was dried
under high vacuum yielding 1.232 g 6b (62% based on 1).
Ethyl [15N,3-13C]-3-(4-fluoroanilino)-2-
13
17
(2,4,5-trifluorobenzoyl)-acrylate (4b, C CH13F145NO3)
A mixture of 1.210 g ethyl 2,4,5-trifluorobenzoylacetate, (1;
4.917 mmol), 1.064 g [formyl-13C]-triethyl orthoformate
123

Synthesis of [1-15N,2-13C]-labeled difloxacin
1227
3. Brown SA (1996) J Vet Pharmacol Ther 19:1
4. Golet EM, Xifra L, Siegrist H, Alder AC, Giger W (2003)
Environ Sci Technol 37:3243
5. Thiele-Bruhn S (2003) J Plant Nutr Soil Sci 166:145–167
6. Boxall ABA, Fogg LA, Blackwell PA, Kay P, Pemberton EJ,
Croxford A (2004) Rev Environ Contam Toxicol 180:1
7. Ku¨mmerer K (2003) J Antimicrob Chemother 52:5
8. Halling-Sørensen B, Nielsen SN, Lanzky PF, Ingersiev F, Holten
Lu¨tzhoft HC, Jørgensen SE (1998) Chemosphere 36:357
The product was analyzed by TLC (solvent system I:
Rf = 0.19; solvent system II: Rf = 0.35).
[1-15N,2-13C]-Difloxacin {[1-15N,2-13C]-1-
(4-fluorophenyl)-6-fluoro-7-(4-methyl-1-piperazinyl)-
1,4-dihydro-4-oxoquinoline-3-carboxylate}
(8b, C CH19F2N125NO3)
13
20
Intermediate 6b was put under argon and dissolved in
30 cm3 of dry acetonitrile. After addition of 3.0 cm3 of
triethylamine and 1.537 g of 1-methylpiperazine (7;
1.7 cm3, 15.345 mmol), the reaction mixture was refluxed
for 300 min and cooled to 22 °C. The mixture was kept at
4 °C over night, and the pearly to light sandy product was
subsequently separated by filtration, washed with acetoni-
trile, and dried under high vacuum. Yield of 8b was 1.225 g
(62% based on 1). The product 8b was examined by TLC
¨
9. Kreuzig R, Holtge S, Brunotte J, Berenzen N, Wogram J, Schulz
R (2005) Environ Toxicol Chem 24:777
¨
10. Kreuzig R (2002) Mitt Umweltchem Okotox 8:9
¨
11. Kreuzig R, Heise J, Holtge S (2006) Mitt Umweltchem Okotox
¨
12:39
12. Schmidt B, Ebert J, Lamshoft M, Thiede B, Schumacher-Buffel
¨
¨
R, Ji R, Corvini PFX, Schaffer A (2008) J Environ Sci Health
Part B 43:8
¨
13. Lamshoft M, Sukul P, Zuhlke S, Spiteller M (2007) Anal Bioanal
¨
Chem 388:1733
¨
14. Schmidt B, Schaffer A, unpublished results
15. Dec J, Bollag JM (1997) Soil Sci 162:858
1
(solvent system II: Rf = 0.23), HPLC, H, 13C, and 15N
NMR. 1H NMR (600 MHz, trifluoroacetic acid-d1): the
spectrum was similar to that of the non-labeled compound,
except the signal of hydrogen atom H-2, which appeared as a
doublet (J = 190.8 Hz, 1H–13C coupling) at d = 9.37 ppm;
13C NMR (151 MHz, trifluoroacetic acid-d1): the spectrum
was similar to that of the non-labeled compound, except the
large signal of carbon atom C2, which appeared as a doublet
(J = 42 Hz, 13C–15N coupling) at d = 152 ppm; 15N NMR
(61 MHz, trifluoroacetic acid-d1): doublet (J = 38.4 Hz,
15N–13C coupling) at d = -215.7 ppm.
16. Chu DTW, Fernandes PB, Claiborne AK, Pihulaec E, Nordeen
CW, Maleczka RE, Pernet AG (1985) J Med Chem 28:1558
17. Chu DTW, Fernandes PB, Maleczka RE, Nordeen CW, Pernet
AG (1987) J Med Chem 30:504
18. Glushkov RG, Marchenko NB, Levshin IB (1997) Pharm Chem J
31:240
19. Baker WR, Cai S, Dimitroff M, Fang L, Huh KK, Ryckman DR,
Shang X, Shawar RM, Therrien JH (2004) J Med Chem 47:4693
20. Grohe K, Heitzer H (1987) Liebigs Ann Chem 1987:29
21. Granneman GR, Snyder KM, Shu VS (1986) Antimicrob Agents
Chemother 30:689
22. Garcia MA, Solans C, Aramayona JJ, Rueda S, Bregante MA
(2000) Chromatographia 51:487
23. Freedom of Information Office, Center for Veterinary Medicine,
FDA (1997) Freedom of information summary, NADA 141-096,
difloxacin hydrochloride. http://www.fda.gov/cvm/FOI/893.htm.
Last accessed 21 May 2009
24. Wetzstein HG, Schmeer N, Karl W (1997) Appl Environ
Microbiol 63:4272
Acknowledgments We are grateful to Prof. Dr. W. Leitner, Dr. J.
Klankermayer, Dr. F. H. Kowaldt, T. Pullmann, and I. Bachmann-
Remy (Institute for Technical and Macromolecular Chemistry,
RWTH Aachen University) for support with synthesis and NMR
spectra. The study was funded by the German Research Foundation,
DFG (Research Unit FOR566 ‘Veterinary Medicines in Soils: Basic
Research for Risk Analysis’).
25. Wetzstein HG, Stadler M, Tichy HV, Dalhoff A, Karl W (1999)
Appl Environ Microbiol 65:1556
26. Choma IM (2003) J Liq Chrom Rel Technol 26:2673
27. Wallis SC, Gahan LR, Charles BG, Hambley TW (1994) Aust J
Chem 47:799
References
28. Holzgrabe U, Branch SK (1994) Magn Reson Chem 32:192
29. Spragg RA (1968) J Chem Soc B:1128
1. Wolfson JS, Hooper DC (1989) Clin Microbiol Rev 2:378
2. Shryock TR (2001) J Clin Microbiol 39:1680
123