Synthesis, absolute configuration and intermediates of 9-fluoro- 6,7- dihydro-5-methyl-1-oxo-1H,5H-benzo[i.j]quinolizine-2-carboxylic acid (flumequine)

Article abstract of DOI:10.1016/S0957-4166(99)00071-3

The antibacterial agent 9-fluoro-6,7-dihydro-5-methyl-1-oxo-1H,5H- benzo[i.j]quinolizine-2-carboxylic acid (flumequine) was synthesized in optically active form from 6-fluoro-2-methyl-1,2,3,4-tetrahydroquinoline (FTHQ). Racemic FTHQ was resolved with the enantiomers of 3-bromocamphor-8- sulfonic acid. The configurations were established by X-ray structures of the two diastereoisomeric salts. Enantiomeric excesses were determined by ¹H NMR analysis.

Full text of DOI:10.1016/S0957-4166(99)00071-3

Pergamon
Tetrahedron: Asymmetry 10 (1999) 1079–1087
Synthesis, absolute configuration and intermediates of 9-fluoro-
6,7-dihydro-5-methyl-1-oxo-1H,5H-benzo[i,j]quinolizine-
2-carboxylic acid (flumequine)
József Bálint,a Gabriella Egri,a,∗ Elemér Fogassy,a Zsolt Böcskei,b Kálmán Simon,b
Antal Gajáry b and Antal Friesz b
aDepartment of Organic Chemical Technology, Technical University Budapest, P.O. Box 91, Budapest H-1521, Hungary
bChinoin Pharmaceutical and Chemical Works Ltd, P.O. Box 110, Budapest H-1325, Hungary
Received 1 February 1999; accepted 25 February 1999
Abstract
The antibacterial agent 9-fluoro-6,7-dihydro-5-methyl-1-oxo-1H,5H-benzo[i,j]quinolizine-2-carboxylic acid
(flumequine) was synthesized in optically active form from 6-fluoro-2-methyl-1,2,3,4-tetrahydroquinoline(FTHQ).
Racemic FTHQ was resolved with the enantiomers of 3-bromocamphor-8-sulfonic acid. The configurations were
established by X-ray structures of the two diastereoisomeric salts. Enantiomeric excesses were determined by 1H
NMR analysis. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
9-Fluoro-6,7-dihydro-5-methyl-1-oxo-1H,5H-benzo[i,j]quinolizine-2-carboxylic acid (flumequine, 4,
Fig. 1) is an antibacterial agent, having a significant difference between the effectiveness of the two
enantiomers. Thus, a convenient and cost-effective method for preparing the isomers is of interest.
One method is described in the literature: Rohlfing et al. performed the optical resolution of the
intermediate 6-fluoro-2-methyl-1,2,3,4-tetrahydroquinoline (FTHQ, 1) by separating its enantiomers via
Figure 1. 9-Fluoro-6,7-dihydro-5-methyl-1-oxo-1H,5H-benzo[i,j]quinolizine-2-carboxylic acid (flumequine, 4)
∗
Corresponding author. E-mail: egri.oct@chem.bme.hu
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00071-3

1080
J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
their diastereoisomeric amides of N-tosyl-(S)-proline1 and synthesized optically active flumequine from
optically active FTHQ. However, this method involves covalently bonded diastereoisomers, which makes
the process impractical and complicated. Here we report our results on the resolution of FTHQ using a
simple diastereoisomeric salt formation method.
2. Results and discussion
Having failed to separate the enantiomers of racemic flumequine we focused on resolving the
FTHQ intermediate and manufacturing optically active flumequine from it. 2-Methyl-1,2,3,4-
tetrahydroquinoline and 2,6-dimethyl-1,2,3,4-tetrahydroquinoline have previously been resolved by
3-bromocamphor-8-sulfonic acid.2,3 This fact, based on the observation that similar molecules can
often be resolved with the same resolution agent,4 prompted us to try this resolving agent first. The
3-bromocamphor-8-sulfonic acid enantiomers proved to be adequate when applied in a molar ratio of
1:1 [represented on the example of the (R)-isomer, Scheme 1].
Scheme 1. Resolution of FTHQ (rac-1)
Racemic FTHQ was dissolved in hot hydrochloric acid solution and the resolution agent was added
as its ammonium salt. Hydrochloric acid excess was necessary to prevent dissociation of the weakly
basic FTHQ. The crystalline phase was enriched in the (R)-1·(R)-BRC8, and the mother liquor in the
(S)-1·(R)-BRC8 diastereoisomeric salt. The solid diastereoisomeric salt gave enantiomerically pure (R)-
1 after cleavage following two recrystallization steps from 5% HCl solution. Enantiomerically pure (S)-1
was obtained using (1S)-3-bromocamphor-8-sulfonic acid as described above.
For production of racemic flumequine from racemic FTHQ, the following method5 was applied
(Scheme 2).
We carried out the above reaction route starting from both the racemate and enantiomerically pure
1. The results are summarized in Table 1. As the reactions do not affect the stereogenic centre, (R)-
flumequine can be obtained from (R)-FTHQ essentially without racemization. Similarly, (S)-FTHQ gives
(S)-flumequine.
X-Ray crystallography established the absolute configurations of both S-1·(R)-BRC8 and R-1·(R)-
BRC8 salts (Figs. 2 and 3). In addition, it has also revealed the possible structural basis of the successful
resolution of racemic FTHQ with 3-bromocamphor-8-sulfonic acid. The structure of S-1·(R)-BRC8
is hydrogen bond donor deficient. The ratio of the hydrogens to acceptor pairs is 2:8. The hydrogen
bonds form endless chains travelling through the crystal structure (Fig. 4). Only two of the oxygens of
the SO−3 group are able to participate in hydrogen bonds and the hydrogen bond chains are based on
HNH+. . .OSO2−. . .HNH+. . .OSO−2 repeating units. These hydrogen bond chains are not interconnected.
The carbonyl group of the camphorsulfonic acid is deeply embedded in a hydrophobic layer of the crystal
structure.

J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
1081
Scheme 2. Synthesis of racemic flumequine
Table 1
Synthesis of flumequine from FTHQ (racemic and optically active routes)

1082
J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
Figure 2. Structure of the ion pair as found in the crystal of S-1·(R)-BRC8
Figure 3. Structure of the ion pair as found in the crystal of R-1·(R)-BRC8
On the contrary, in R-1·(R)-BRC8 two water molecules are also included in the crystal lattice. The
ratio of acidic hydrogens to acceptor pairs is only 6:12. Each oxygen of the SO−3 groups participates
in hydrogen bonding. The hydrogen bonds form two-dimensional networks utilizing the two water
molecules (as mediators) as well as all three oxygen atoms of the sulfonate moiety. In addition, the
carbonyl group of the camphorsulfonate is also involved in the hydrogen bond network since one of the
water molecules sinks deeply into the hydrophobic layer to form a hydrogen bond with that carbonyl
group (Fig. 5). The more extended hydrogen bond network — which is due to the presence of two
additional water molecules — is probably responsible for the higher stability of the R-1·(R)-BRC8 salt.

J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
1083
Figure 4. Crystal packing diagram of S-1·(R)-BRC8
3. Experimental
3.1. Materials and methods
The 1H NMR spectra were recorded at 500 MHz on a Bruker DRX-500 spectrometer. Chemical shift
values are expressed in ppm values on the δ scale. IR spectra of thin film samples were taken on a
Specord 2000 spectrometer. Optical rotations were determined on a Perkin–Elmer 241 polarimeter. (S)-
and (R)-3-Bromocamphor-8-sulfonic acid ammonium salts were purchased from Aldrich. FTHQ was
manufactured in Chinoin. All solvents were freshly distilled.
3.2. Preparation of 2 from FTHQ
Racemic FTHQ (1.0 g, 6.05 mmol, mp 34–36°C) was stirred with diethylethoxymethylene malonate
(1.3 g, 6.01 mmol) at 125°C for 2.5 h. The mixture was kept at this temperature in vacuo for 1 h to yield
a brownish yellow, viscous oil (1.97 g, 5.87 mmol, 97.0%).
The same reaction starting from (S)-FTHQ ([α]D=−70.0, c=1, ethanol) yielded brownish yellow
viscous oil (97.0%, [α]D=+264.7, c=1, ethanol; +287.9, c=1, dimethyl formamide).
The same reaction starting from (R)-FTHQ ([α]D=+70.2, c=1 ethanol) yielded brownish yellow
viscous oil (96.1%, [α]D=−266.1, c=1, ethanol; −288.1, c=1, dimethyl formamide).
1H NMR: (CDCl3): 1.21 (d, 3H, CH3), 1.28 (m, 6H, CH3–Et), 1.7–2.1 (ms, 2H, CH2), 2.7–2.9 (ms,
2H, Ar–CH2), 4.0–4.3 (ms, 5H, 2CH2–E+CH), 6.8–7.0 (m, 3H, Ar), 7.7 (s, 1H, _CH). FT-IR (film,

1084
J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
Figure 5. Crystal packing diagram of R-1c·(R)-BRC8
cm−1): 3398, 2981, 2937, 2499, 1695, 1593, 1499, 1438, 1382, 1364, 1291, 1267, 1243, 1227, 1192,
1074, 1045, 1027, 951, 919, 873, 812, 663. Calcd for C18H22NO4F: C, 64.46; H, 6.61; N, 4.18; found C,
64.27; H, 6.60; N, 4.20.
NMR and IR spectra are identical for the racemic and optically active 2.
3.3. Preparation of flumequine ethyl ester 3
Racemic FTHQ (2.0 g, 12.1 mmol, mp 34–36°C) was stirred with diethylethoxymethylene malonate
(3.0 g, 13.9 mmol) at 125°C for 2.5 h. The mixture was kept at this temperature in vacuo (10 mmHg) for
30 min. After cooling, polyphosphoric acid (3.0 g) was added and the mixture was stirred at 110°C for 6
h. After this time, the reaction mixture was poured into water (50 ml) and left to stand for 4–5 h. After
filtration, washing with water (5×10 ml) and drying, 2.98 g of off-white powder was obtained. After
recrystallizing from ethanol, white powder (2.42 g, 8.37 mmol, 69.1%, mp 183–186°C) was obtained.
The same reaction starting from (S)-FTHQ ([α]D=−70.0, c=1, ethanol) yielded white powder (81.9%,
[α]D=−81.5, c=1, ethanol; −73.4, c=1, dimethyl formamide, mp 187–189°C).
The same reaction starting from (R)-FTHQ ([α]D=+70.2, c=1, ethanol) yielded white powder (80.5%,
[α]D=+81.5, c=1, ethanol; +73.9, c=1, dimethyl formamide, mp 188–190°C).
1H NMR: (CDCl3): 1.41 (t, 3H, CH3–Et), 1.49 (d, 3H, CH3), 2.1–2.3 (m, 2H, CH2), 3.0–3.2 (ms, 2H,
Ar–CH2), 4.4 (q, 2H, CH2–Et), 4.5 (m, 1H, CH), 7.2–8.0 (m, 2H, Ar), 8.4 (s, 1H, _CH). FT-IR (KBr,
cm−1): 3420, 2991, 1719, 1623, 1593, 1565, 1486, 1428, 1414, 1311, 1219, 1166, 1136, 1103, 1048, 915,
878, 804, 740, 627. Calcd for C16H16NO3F: C, 66.43; H, 5.57; N, 4.84; found C, 66.23; H, 5.59; N, 4.82.

J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
NMR and IR spectra are identical for the racemic and optically active 3.
3.4. Preparation of flumequine 4
1085
Racemic FTHQ (2.0 g, 12.1 mmol, mp 34–36°C) was stirred with diethylethoxymethylene malonate
(3.0 g, 13.9 mmol) at 125°C for 2.5 h. The mixture was kept at this temperature in vacuo (10 mmHg)
for 30 min. After cooling, polyphosphoric acid (3.0 g) was added and the mixture was stirred at 110°C
for 6 h. After adding water (6 ml), it was stirred at 100°C for 3 h and then 1 h at 0°C, then filtered.
The cake was washed with water (5×2 ml), ethanol (3×1 ml) and was dried to yield 2.47 g of yellowish
material, which was then dissolved in hot dimethyl formamide (5 ml), cooled to 0–3°C, filtered, washed
with dimethyl formamide (2×0.5 ml), water (3×2 ml), ethanol (2×1 ml) and dried to yield white crystals
(2.24 g, 8.57 mmol, 70.8%, mp 256–259°C).
The same reaction starting from (S)-FTHQ ([α]D=−70.0, c=1 ethanol, mp 40–42°C) yielded white
crystals (77.5%, [α]D=−48.6, c=0.21, dimethyl formamide, mp 246–249°C).
The same reaction starting from (R)-FTHQ ([α]D=+70.2, c=1, ethanol, mp 40–43°C) yielded white
crystals (75.4%, [α]D=+48.9, c=0.21, dimethyl formamide, mp 247–250°C).
1H NMR: (CDCl3): 1.55 (d, 3H, CH3), 2.2–2.3 (m, 2H, CH2), 3.0–3.2 (m, 2H, Ar–CH2), 4.6 (m, 1H,
CH), 7.3–8.0 (m, 2H, Ar), 8.7 (s, 1H, _CH), 14.9 (s, 1H, COOH). FT-IR (KBr, cm−1): 3448, 3055, 2990,
2723, 1724, 1621, 1567, 1518, 1471, 1423, 1394, 1376, 1340, 1318, 1303, 1272, 1212, 1196, 1134, 1067,
974, 899, 886, 810, 719, 698, 646. Calcd for C14H12NO3F: C, 64.36; H, 4.63; N, 5.36; found C, 64.55;
H, 4.64; N, 5.36.
NMR and IR spectra are identical for the racemic and optically active 4.
3.5. Optical resolution of FTHQ 1
Racemic FTHQ (2.0 g, 12.1 mmol) and (R)-3-bromocamphor-8-sulfonic acid ammonium salt (4.0
g, 12.2 mmol) were dissolved in hot hydrochloric acid solution (1.5 ml of concentrated hydrochloric
acid and 20 ml of water). The mixture was cooled to room temperature. Crystallization was initiated
by scratching. After standing at room temperature with occasional stirring for 1 h it was filtered. The
diastereoisomeric salt was recrystallized twice from 5% hydrochloric acid solution (5 ml). After filtration
and drying, white crystalline material was obtained (1.71 g, [α]D=+86.3, c=1, methanol, mp 169–172°C).
Ethyl acetate (20 ml) was added and the mixture was washed with 1 M Na2CO3 (3×10 ml). The
organic phase was dried over Na2SO4 and the solvent was evaporated in vacuo to yield white crystals
(0.52 g, [α]D=+70.0, c=1, ethanol, mp 40–42°C).
For the mother liquor, ethyl acetate (20 ml) was added and the mixture was washed with 1 M Na2CO3
(3×30 ml). The organic phase was dried over Na2SO4 and the solvent was evaporated in vacuo to yield
an oily-crystalline material (1.10 g, [α]D=−33.4, c=1, ethanol).
The other enantiomer of FTHQ was obtained similarly with the enantiomer from the resolving agent.
3.6. X-Ray crystallography
Crystals were obtained as follows.
(S)-(−)-FTHQ (0.20 g, 1.21 mmol, [α]D20=−70, c=1, ethanol) and (R)-3-bromocamphor-8-sulfonic
acid monohydrate (0.42 g, 1.27 mmol) were dissolved in hot water (5 ml). The solution was cooled
to room temperature, inoculated and left to crystallize overnight. The crystals were filtered and air-

1086
J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
Table 2
Crystal data and experimental details
dried to yield white needles (0.40 g), [α]D20=+30.8 (c=1, methanol), mp: 178–181°C. Calcd for
C20H27BrFNO4S: C. 50.42; H, 5.71; N, 2.94; found C, 50.31; H, 5.72; N, 2.95.
(R)-(+)-FTHQ (0.20 g, 1.21 mmol, [α]D20=+70.2, c=1, ethanol) and (R)-3-bromocamphor-8-sulfonic
acid monohydrate (0.42 g, 1.27 mmol, [α]D,dry mat.=+105 to +108, c=1, water) were dissolved in hot
water (6 ml). The solution was cooled to room temperature, inoculated and left to crystallize for 1 h. The
crystals were filtered and air-dried to yield white crystals (0.46 g), [α]D20=+86.5 (c=1, methanol), mp:

J. Bálint et al. / Tetrahedron: Asymmetry 10 (1999) 1079–1087
1087
at 78–82°C partially dissolved in its own crystal water, above 100°C lost crystal water and finally melted
at 171–174°C. Calcd for C20H31BrFNO6S: C, 46.88; H, 6.10; N, 2.73; found C, 47.05; H, 6.12; N, 2.72.
Data were recorded at 293 K with a Rigaku AFC6S diffractometer using graphite monochromatized
CuKα. Three intensity standards were measured every 150 reflections and indicated no decay during
data collection. Data reduction included corrections for Lorentz and polarization effects.
The structure was solved using direct methods. Coordinates and anisotropic thermal motion parameters
for all non-hydrogen atoms were refined. Hydrogen atoms were placed in calculated positions depending
on the nature of the atom bonded to the hydrogen. Isotropic thermal motion parameters were refined by
constraining chemically similar hydrogen atoms to obtain the same values.
The programs used for absorption corrections, data reduction, structure solution and presentation were
from the TEXSAN package.6 Refinement was performed using SHELXL-93.7 The results are shown in
Table 2.
Acknowledgements
The Hungarian OTKA Foundation (Project Nos. 1759 and 14894) is gratefully acknowledged for
financial support. J. B. thanks the Zoltán Magyary Foundation for a postdoctoral fellowship. G. E. thanks
the OTKA foundation for a postdoctoral fellowship (D29445).
References
1. Gerster, J. F.; Rohlfing, S. R.; Winady, R. M. North American Medicinal Chemistry Symposium; Toronto, 1982.
2. Pope, W. J.; Peachey, S. J. J. Chem. Soc. 1899, 75, 1066–1093.
3. Pope, W. J.; Rich, E. M. J. Chem. Soc. 1899, 75, 1093–1105.
4. Gizur, T.; Péteri, I.; Harsányi, K.; Fogassy, E. Tetrahedron: Asymmetry 1996, 6, 1589–1590.
5. Patent: DE 2,264,163, 1973; CA 79:92029y, 1973.
6. Molecular Structure Co. TEXSAN, Single Crystal Structure Analysis Software; MSC, 3200 Research Forest Drive, The
Woodlands, TX 77381, USA, 1992.
7. Sheldrick, G. M. SHELXL-93, Program for the Refinement of Crystal Structures; The University of Göttingen, Göttingen,
Germany, 1993.