Chiral DNA gyrase inhibitors. 3. Probing the chiral preference of the active site of DNA gyrase. Synthesis of 10-fluoro-6-methyl-6,7-dihydro-9- piperazinyl-2H-benzo[a]quinolizin-20-one-3-carboxylic acid analogues

Article abstract of DOI:10.1021/jm0401356

In pursuit of an apparent literature anomaly, S- and R-6-methyl-6,7- dihydro-2H-benzo[a]-quinolizin-2-one-3-carboxylic acids (12 and 22) were synthesized by an unambiguous route from optically active norephedrines, and their antibacterial potencies were measured. Against Gram-negative microorganisms and DNA gyrase a preference for S-absolute configuration was found whereas R-absolute stereochemistry was more active against Gram-positives. These results are in partial conflict with an earlier report. In an attempt to enhance potency, racemic 10-fluoro-9-piperazinyl (35) and related analogues were synthesized by a novel route. The latter analogues were surprisingly unimproved in potency. The implications of these findings are briefly discussed.

Full text of DOI:10.1021/jm0401356

J. Med. Chem. 2005, 48, 1229-1236
1229
Chiral DNA Gyrase Inhibitors. 3. Probing the Chiral Preference of the Active
Site of DNA Gyrase. Synthesis of 10-Fluoro-6-methyl-6,7-dihydro-9-piperazinyl-
2H-benzo[a]quinolizin-20-one-3-carboxylic Acid Analogues
Robert A. Fecik,† Pratik Devasthale,† Segaran Pillai,† Ali Keschavarz-Shokri,† Linus Shen,‡ and
Lester A. Mitscher*,†
Departments of Medicinal Chemistry and of Molecular Biosciences, University of Kansas, 4010 Malott Hall,
1251 Wescoe Hall Drive, Lawrence, Kansas 66045-7582, and Antiinfectives Research Department, Abbott Laboratories,
Abbott Park, Illinois 60064
Received July 16, 2004
In pursuit of an apparent literature anomaly, S- and R-6-methyl-6,7-dihydro-2H-benzo[a]-
quinolizin-2-one-3-carboxylic acids (12 and 22) were synthesized by an unambiguous route from
optically active norephedrines, and their antibacterial potencies were measured. Against Gram-
negative microorganisms and DNA gyrase a preference for S-absolute configuration was found
whereas R-absolute stereochemistry was more active against Gram-positives. These results
are in partial conflict with an earlier report. In an attempt to enhance potency, racemic 10-
fluoro-9-piperazinyl (35) and related analogues were synthesized by a novel route. The latter
analogues were surprisingly unimproved in potency. The implications of these findings are
briefly discussed.
Introduction
bacteria and at the enzyme level than the distomeric
R-enantiomers.
The empirically driven evolution of ever more effective
quinolone antiinfective agents is a powerful tribute to
the power of modern medicinal chemical methodology.
Nonetheless, a detailed understanding of their molec-
ular modes of action and toxicity would be even more
efficient. Unfortunately, accomplishing this presents
major present challenges. These drugs are now known
to operate through more than one target and target
components. DNA gyrase, topoisomerase IV, and, to a
lesser extent, human topoisomerase II all have signifi-
cantly relevant affinities for quinolones when in the
presence of target DNA. None of these enzymes have
been crystallized. Further, the binding site consists of
enzyme, DNA, and up to four quinolone molecules.
Capturing all this in a crystal suitable for X-ray analysis
will be a major accomplishment.
At present we are forced to rely instead on insights
gained from analysis of the utility of specific quinolone
structures in interfering with this process. The underly-
ing assumption, however reasonable, is that all of these
agents are acting in precisely the same manner.
With this backdrop in mind, in this paper we probe
and examine further through synthesis of chiral ligands
the best presently accepted predictive model for qui-
nolone action, the Shen model as extended by Morressey
et al. This model provides a satisfactory rationale for
the very significant eudismic ratio observed when
substituents attached to the quinolone nitrogen atom
contain nearby chiral centers. Starting with flumequine
and continuing with levofloxacin and subsequently with
many more examples, it has been established that the
S-enantiomers are very much more active in intact
There is, however, a preliminary report, albeit involv-
ing a different ring system, in which the opposite
stereochemical preference is asserted. Operating on the
principle that it is the exception that probes the rule
and when properly interpreted can deepen our under-
standing, we have repeated this work and addressed
this anomaly.
The absolute stereochemistry of a substituent con-
taining a nearby chiral center attached near N-1 of rigid
analogues of tricyclic quinolone antiinfective agents
produces a very significant eudismic effect when the
absolute chemistry is S in the cases of flumequine (1),1
S-25930 (2),1 benofloxacin (3),2 and nadifloxacin (4).3
The bioisosteric replacement of O for methylene as in
levofloxacin (5)4,5 and S-12681 (6),6 as well as bioiso-
electronic replacement of S for methylene as in GRB-
23790 (7)7 and WIN-58161 (8),8 produces the same
biological preference. The enantiopreference for S-
configuration has now become orthodox as exemplified
by recent development candidates pazufloxacin (T-3761)
(9)9 and DV-7751a (10)10 and many other lead mol-
ecules. These structures are illustrated in Figure 1. This
differential effect of chirality has been rationalized to
be due to asymmetric stacking in a cooperative model
of the molecular mode of action of these drugs.11
It is known also that the nature of the substituent
attached to the R-carbon is significant for binding to
DNA gyrase as opposed to human topoisomerase II.12
Against the latter enzyme (associated with certain
cytotoxicities), compounds with a methyl group are less
active than those with a hydrogen atom or a methylene
moiety at the corresponding position.
* To whom correspondence should be addressed. Phone: (785) 864-
4562. Fax: (785) 864-5326. E-mail: lmitscher@ku.edu.
† University of Kansas.
Thus, our attention was drawn to a preliminary
literature report that in a different ring system (the
6-methyl-6,7-dihydro-2H-benzo[a]quinolizin-2-one-3-car-
‡ Abbott Laboratories.
10.1021/jm0401356 CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/26/2005

1230 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 4
Fecik et al.
N-acetyl group was hydrolyzed to produce chiral amine
16, isolated in 90% yield as its hydrochloride salt. This
was then formylated to amide 17 in 62% yield by
heating with ethyl formate. A Bischler-Napieralski
cyclization of 17 using polyphosphoric acid gave chiral
3,4-dihydroisoquinoline 18 in 65% yield as its hydro-
chloride salt.
Reaction of ethyl acetoacetate with acetic anhydride
and trimethyl orthoformate followed by reaction with
diethylamine produced Michael acceptor diethylami-
nomethylene ethyl acetoacetate (19).15 This cyclized to
a 2:1 mixture of diastereoisomers (20) in 67% yield upon
condensation with 18. Oxidation of the resulting di-
astereoisomeric mixture with chloranil produced ester
21 in 82% yield. Sodium hydroxide hydrolysis produced
24
the desired optically active target substance 12 ([R]D
+189.4° in chloroform) in 82% yield. This nine-step
sequence of reactions gave overall a 10% yield.
Figure 1. Structures of many of the previously described
quinolone antiinfective agents possessing chiral substituents
attached to N-1.
The S-enantiomer 22 (Ro-14-5319) ([R]D24 -185.1° in
chloroform) was prepared with equivalent efficiency
following the same route but starting with 1R,2S-(-)-
norephedrine.
boxylic acids such as Ro-14-5319 (11) and its des-
methylenedioxo analogue Ro-14-4299) (12)) the eutomer
is reported to be R.13,14 To advance our understanding
of the structural requirements of the drug-binding
pocket for quinolones, we decided to explore this seem-
ing anomaly further by preparation of analogues using
starting materials from the chiral pool so that the
absolute stereochemistry of the products would be
unambiguously established and additional chiral ana-
logues could be prepared efficiently.
Considering the generalization based upon Pfeiffer’s
observation that chiral recognition most often increases
in parallel with potency and/or receptor affinity16 the
synthesis of 10-fluoro-6-methyl-6,7-dihydro-9-piperazi-
nyl-2H-benzo[a]quinolizin-20-one-3-carboxylic acid (35)
was undertaken next (Scheme 2). The electronic con-
tributions of the new groups to the pharmacophoric
pyridine ring would be similar to those found in cipro-
floxacin and levofloxacin. Synthesis of the necessary
analogues by the same route as used for 12 and 22 failed
due to the deactivating influences of the required
fluorine-containing aromatic ring substituents in the
intermediates, and a new route was therefore devised.
This route began with commercially available 3,4-
difluorobenzaldehyde (23) and was carried out initially
with racemic materials with the intention of later
repeating the work with optically active materials or
by resolution by any of a number of means and, in that
case, identifying the stereochemistry of the eutomer and
distomer by chiroptical comparisons with 12 and 22.
Thus, 3,4-difluorobenzaldehyde 23 was subjected to a
Henry condensation by heating with nitroethane and
ammonium acetate to produce nitro alkene 24 in 94%
yield, and this in turn was converted to racemic primary
amine 25 in 91% yield upon lithium aluminum hydride
reduction. Formation of formamide 26 by reaction with
Chemistry
For the record, since the experimental details of the
earlier work have not been published, we report herein
the details of the synthesis of (R)-Ro-14-4299 (12)
starting with different, optically active materials. This
route (Scheme 1) proceeded without significant difficulty
in our hands after only minor modifications following
the briefly reported reaction sequence.14 Because direct
hydrogenolysis of the unneeded benzylic hydroxyl group
of (1S,2R)-(+)-norephedrine (13) required rather forcing
conditions, it was first diacetylated to ester amide (14)
in 92% yield with acetic anhydride-pyridine. Hydro-
genolysis of 14 then proceeded smoothly at atmospheric
pressure and 80-90° C using 5% palladium-barium
sulfate in 70% perchloric acid solution to produce R-2-
acetamino-1-phenylpropane (15) in 72% yield. Next the
Scheme 1. Synthesis of R- and S-6-Methyl-6,7-dihydro-2H-benzo[a]quinolizin-2-one-3-carboxylic Acids

Chiral DNA Gyrase Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 4 1231
Scheme 2. Synthesis of 10-Fluoro-6-methyl-6,7-
dihydro-9-piperazinyl-2H-benzo[a]quinolizine-20-one-
3-carboxylic Acid (35) and Its 9-Dimethylhydrazino
Analogue (36).
in a hetero Diels-Alder reaction and that reactions of
Danishefsky’s diene and related dienes with cyclic
imines were known to proceed well in the synthesis of
a variety of alkaloids,18 this reaction was next at-
tempted.
Reaction of ethyl ethoxymethyleneacetoacetate (30)
with trimethylsilyl chloride and zinc chloride-triethyl-
amine produced the carbethoxy analogue of the Dan-
ishefsky-type diene (31) in 75% yield.
When diene 31 was reacted with cyclic imine 28, the
desired dihydropyridone (32) was formed smoothly as
a mixture of diastereoisomers that proved difficult to
characterize spectroscopically, so it was aromatized
promptly with chloranil to give readily characterized
pyridone 33 in 47% yield for the two steps. The
synthesis of the desired fluoroquinolone isomer was
completed by alkaline hydrolysis to the corresponding
acid (34) followed by a nucleophilic aromatic displace-
ment reaction with piperazine to give the racemic target
compound 35.
The structure assigned to compound 35 is consistent
with the normal outcome of analogous reactions in the
well precedented benzopyridone parent series. In that
case, the fluorine at position 7 is activated toward
displacement over that at C-6 by the electron-withdraw-
ing character of the carbonyl at C-4. As a result, the
reaction produces products such as 1-10, ciprofloxacin,
norfloxacin, and a host of others. Clearly the same
directing influences are expected to operate in the
benzo[a]quinolizinone ring system of compound 34 due
to the inductive effect of the analogous keto group.
For the reasons explicated in the Discussion section,
analogue 36 was prepared in 64% yield by a boron
difluoride-assisted nucleophilic aromatic displacement
reaction with 1,1-dimethylhydrazine.
acetic formic anhydride proceeded in 100% yield. In
contrast to the result when the aromatic ring contained
only hydrogens, cyclization of 26 under a wide variety
of standard Bischler-Napieralski conditions failed vir-
tually completely. One presumes that the cyclization of
the Ritter-type intermediate in the normally preferred
manner is frustrated by decreased electronic contribu-
tions from the difluorinated aromatic ring.
Ultimately the modified Bischler-Napieralski condi-
tions developed by Larsen et al. were found to succeed
admirably despite their report that this variant gives
poor results with electron deficient reactants.17 Reaction
of 26 with oxalyl chloride followed by cyclization with
iron(III) chloride produced the diastereoisomeric oxalyl
ester amide 27 in 85% yield for the two steps. The
success of this process is likely due to the enhanced
nucleophilicity of the presumed cyclic oxazolidinedione
intermediate which forms a highly reactive N-acylimin-
ium ion (29) further polarized in the presence of iron-
(III) chloride.
Results and Discussion
The R- and S-enantiomers 12 and 22 were tested in
vitro against several Gram-positive and Gram-negative
bacteria as well as against purified DNA gyrase isolated
from Gram-negative Escherichia coli and Gram-positive
Micrococcus luteus using standard, published, method-
ology.15 The results are recorded in Table 1. In the
present work these enantiomers are considerably less
potent than has been reported for Ro-14-9578 enanti-
omers. Nevertheless the S-enantiomer was the more
potent against the purified enzyme from E. coli H560
and against two strains of intact E. coli (SS [supersensi-
tive] and H560 [from which the enzyme preparation was
derived]. Against the Juhl strain, however, there was
no enantiopreference, and against the KNK 437, strain
potency was too weak to decide. Against Pseudomonas
aeruginosa, once again the S-enantiomer was preferred.
Thus, against intact Gram-negative microorganisms or
against the purified DNA gyrase from E. coli the
S-enantiomer (22) shows the same enantiopreference as
does 1-10 so that 12 and 22 are not exceptional in this
regard. These results conflict with those reported in the
meeting abstract and poster.
Compound 27 was immediately hydrolyzed and de-
hydrated with acidic methanol to give the previously
elusive 28 in 73% yield. Unfortunately, however, this
intermediate did not undergo the tandem Michael
addition/cycloalkylation reaction with enamine 19 that
had taken place previously with its nonhalogenated
analogue (18).
The bioactivity against Gram-positive microorganisms
E. facium, and M. luteus and from DNA gyrase obtained
from it was too low to lead to any conclusion. On the
other hand, against S. aureus the R-enantiomer was
preferred in accord with the literature report for 11 and
Considering that the electron deficient imine moiety
of dihydroisoquinoline 28 would likely be advantageous

1232 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 4
Fecik et al.
Table 1. In Vitro Minimum Inhibitory Concentration Values
(MIC, µg/mL) and Enzyme Inhibitory Concentrations for
Synthetic Analogues Compared to Standard Ciprofloxacin (38)
same spatial relationship with respect to the keto acid
moiety of typical fluoroquinolones. It was no more active
than 35 making this rationalization unlikely. Upon the
basis of these disappointments, synthesis of additional
analogues of 12 and 22 and of the enantiomers of 35
and 36 was not undertaken.
microorganism
12
22
35
25
36
38
Staphylococcus aureus
ATCC 6538P
Micrococcus luteus 9341
Enterobacterium faecium
ATCC 8043
Escherichia coli Juhl
Escherichia coli SS
Escherichia coli H560
25
50
25 0.2
100
>100
>100
>100
-
-
-
-
1.56
0.78
Conclusions
25
25
1.56
50
-
-
0.2
50
30
25 0.02
Many publications support the idea that a in tricyclic
quinolone analogues containing a methyl group on an
asymmetrically substituted carbon attached to N-1, as
with ofloxacin and its analogues, superior antimicrobial
activity is obtained when the absolute stereochemistry
is S. A single report challenges this in molecules
possessing a different ring system. The results of this
study do not agree with these conclusions when Gram-
negative microorganisms are involved. Against Gram-
positives, however, our results agree that the eutomer
is R. Surprisingly, adornment of these new analogues
with substituents associated with high potency among
other quinolones failed to result in an improvement.
Reorienting the distal nitrogen atom so that it can
occupy the same space as that in classical quinolones
also failed to result in an improvement. It seems likely
then that the active site of the various enzymatic targets
for these drugs differs sufficiently to distinguish be-
tween these analogues. Clearly, more work will be
required to settle this issue conclusively. When a
modeling consensus emerges that rationalizes the po-
tency of fluoroquinolones at the molecular level, it
should have sufficient power to rationalize these new
findings.
6.25
-
-
-
0.005
0.01
0.2
50
25
>100
50
Escherichia coli KNK 437 >100
Pseudomonas aeruginosa
E. coli H560 DNA gyrase
(IC50 in µg/mL)a
100
32
>25 0.05
-
20
3.1
M. luteus 934 DNA gyrase >250
>250
-
-
-
(IC50 in µg/mL)a
a The values reported for minimum inhibitor constants were
obtained in the usual 2-fold agar dilution series manner as the
average of triplicate runs. The values for the IC50 and CC50
determinations were obtained as the average of triplicate runs also
using the methodology previously detailed.25 Error bars are rarely
if ever reported in the literature for such measurements; however,
our experience is that the deviation is typically (35% in our hands.
Figure 2. One putative stacking mode of ofloxacin and of 10-
fluoro-6-methyl-6,7-dihydro-9-piperazinyl-2H-benzo[a]quino-
lizin-20-one-3-carboxylic acid.
12. It is possible that the different ring system of the
benzo[a]quinolizines accounts for this partial departure
from the prevailing pattern. Recent work demonstrating
that topoisomerase IV, rather than DNA gyrase, is often
the preferred target for quinolones in Gram-positives
may provide a plausible rationale for this discrepancy.
The eudismic ratios measured for 12 and 22 do not
exceed four against any of the bacteria and is even less
against the purified enzyme. The data reported in the
earlier work14 (utilizing different bacterial strains)
indicated a much higher potency and eudismic ratio
(0.8-6.5 depending upon the strain of E. coli chosen and
>62.5 against S. aureus). A later publication using
racemic Ro-14-5319 (Ro-14-9578) confirmed that the
primary target in Gram-negatives is DNA gyrase.13 No
results were reported for DNA gyrase from Gram-
positives or against bacterial topoisomerase IV (that
enzyme not being known at the time).
It seemed likely that the lower potencies for 12 and
22 were due to the lack of substituents in the aromatic
ring in such as those generally associated with excellent
activity in ordinary fluoroquinolones. Hence, substituted
analogue 35 was prepared and tested. Disconcertingly,
its in vitro antibacterial activity was scarcely improved.
It was initially hypothesized that the relative ar-
rangement in space of the keto acid moieties in ring A
of 12 and 22 result in placement of the distal amino
group of the C-7 piperazino moiety in an incorrect place
compared to that of ordinary quinolones (such as cipro-
floxacin). This idea is illustrated in formula 37 (Figure
2). To test this hypothesis, analogue 36 was prepared
and tested as its distal amino function can occupy the
Experimental Section
Optical rotations were recorded on a Perkin-Elmer pola-
rimeter model 241. The concentrations for optical rotations are
represented in units of g/100 mL. The biological testing was
performed at Abbott laboratories. Melting points were deter-
mined on a Thomas-Hoover capillary melting point apparatus
and are uncorrected. Infrared (IR) spectra were obtained on a
Perkin-Elmer 1420 spectrophotometer. Proton (1H NMR) and
carbon (13C NMR) nuclear magnetic resonance spectra were
recorded on Varian FT-80, Varian XL-300, GE QE-300, or
Bruker AM-500 spectrometers in the indicated solvents using
tetramethylsilane as the internal standard, unless otherwise
specified. All chemical shifts are expressed in parts per million
(δ). Electron impact mass spectra (El), chemical ionization
mass spectra (Cl), and high-resolution mass spectra were
obtained on a Varian CH-5 or Ribermag R10-10 mass spec-
trometer. Microanalyses were performed using a Hewlett-
Packard Model 185 CHN analyzer at the University of Kansas.
Ultraviolet (UV) spectra were recorded on a Hewlett-Packard
Diode Array 8450A. Flash chromatography and medium-
pressure chromatography (MPLC) were performed on Merck
silica gel (230-400 mesh) and gravity column chromatography
on Merck silica gel (70-270 mesh). Preparative centrifugal
thin-layer chromatography (radial chromatography) was per-
formed on a Harrison Model 7924 Chromatotron using Merck
silica gel 60 PF-254.
(1R,2R)-(+)-Norephedrine Diacetate (14). Ac2O (31 mL,
0.33 mol) was added dropwise to a solution of (1S,2R)-(+)-
norephedrine (13, 24.41 g, 0.16 mol) in pyridine (30 mL, 0.37
mol) at r.t. The reaction mixture was stirred for 24 h at r.t.
after which it was poured onto about 10 g of crushed ice.
Keeping the reaction mixture cool, the pH of the solution was
adjusted to about 1 with concentrated HCl. The mixture was
then extracted into CH2Cl2 (100 mL × 3). The organic layer
was washed with 5% NaOH (50 mL × 3) and brine, dried (Na2-
SO4), filtered, and evaporated under reduced pressure to obtain

Chiral DNA Gyrase Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 4 1233
33.4 g (0.142 mol, 88%) of the crude diacetate. Recrystallization
(CH2Cl2/ether/hexanes) afforded 32.5 g (86%) of colorless
crystals: mp 97-98 °C,19 Rf 0.38 (50:50-hexane:ethyl acetate);
1H NMR (80 MHz, CDCl3) δ 1.0 (d, J ) 7.2 Hz, 3H), 1.9 (s,
3H), 2.1 (s, 3H), 4.1-4.6 (m, 1H), 5.75 (d, J ) 4 Hz, 1H), 7.3
(s, 5H); IR (KBr) 3100, 1630, 1540, 1440, 1270, 1150, 1130,
650, 600 cm-1; UV-vis (MeOH) λmax 205 (12 048) nm; MS (El)
CH2CI2 as the solvent system to obtain 1.548 g (62%) of a
colorless liquid. Kugelrohr distillation gave an analytical
sample as a colorless, crystalline solid.: mp 49-51 °C; 1H NMR
(300 MHz, CDCI3) δ 1.1539 (d, J ) 6.66 Hz) and 1.2688 (d, J
) 6.57 Hz, 3H for both), 2.64-2.90 (m, 2H), 4.24-4.44 (m, 1
H), 5.5083 and 5.6910 (both s, 1H for both), 7.10-7.36 (m, 5H),
8.04 and 8.0754 (both s, 1H for both); 13C NMR (75.4 MHz,
CDCI3) δ 19.9533/19.9815, 42.2626, 44.9111/44.9481, 126.5560/
126.8341, 128.4137/128.6525, 129.3855, 137.5160, 160.4354;
24
m/e 236 (M+ + 1, 0.66), 176 (4.9), 86 (99), 44 (100); [R]D
+79.88° (c ) 0.022, CHCl3); Analysis calcd for C13H17NO3: C,
66.36; H, 7.28; N, 5.95; found: C, 66.39; H, 7.60; N, 6.18.
(1S,2S)-(-)-Norepinephrine Diacetate. This compound
was synthesized in 95% yield from 1S,2S-norephedrine fol-
lowing the same procedure. It was recrystallized from CH2-
Cl2/hexane and then ether: mp 97-98 °C. TLC, IR, NMR, and
MS agreed with those of the prior specimen.
[R] 24 +21.07° (c ) 0.3037, CHCl3).24 Analysis calcd for C10H13-
D
N0: C, 73.59; H, 8.03; N, 8.58; found: C, 73.43; H, 8.37; N,
8.70.
(S)-(-)-N-Formylamphetamine. This substance was pre-
pared as above in 50% yield and melted at 49-51 °C following
recrystallization. TLC, IR, NMR, and MS agreed with those
of the prior specimen.
(R)-(+)-Amphetamine Acetate (15). 5% Pd on BaSO4
(5.160 g) was dispersed in glacial HOAc (15 mL) and saturated
with hydrogen at r.t. and atmospheric pressure for about 1.5
h. A mixture of (1S,2R)-(+)-norephedrine diacetate (14, 5.344
g, 22.74 mmol), 70% perchloric acid (5 mL), and glacial HOAc
(20 mL) were added to the catalyst dispersion. The hydrogena-
tion was carried out at 95 °C at atmospheric pressure for 4.5
h. The reaction mixture was then filtered through Celite. The
filtrate was treated with saturated solution of KOH, keeping
the reaction mixture cool (by adding ice) until all potassium
chlorate precipitated out. The solid was filtered off. The filtrate
was further basified with KOH pellets, keeping the mixture
cool (using ice), to pH 10-12. The amphetamine acetate was
extracted into CH2CI2 × 3, and the organic layer was washed
with brine, dried (Na2SO4), filtered, and evaporated in vacuo
to yield 2.866 g (71%) of amphetamine acetate as a colorless
solid. 1.419 g of this product were recrystallized from ether/
hexanes to yield 1.39 g of a colorless, crystalline product: mp
(R)-3-Methyl 3,4-dihydroisoquinoline Hydrochloride
(18). (R)-(+)-N-formylamphetamine (17, 30.5 mg, 0.19 mmol)
was flushed with argon, and hot PPA (1 mL) that was
preheated to 170 °C was added all at once. The reaction flask
was lowered into an oil bath at 170 °C under argon. After 1.5
h, the mixture was poured on ice, concentrated HCl (5 mL)
was added, and neutral components were extracted into ether
(3 × 10 mL). The aqueous layer was basified with 10% NaOH
and the mixture extracted into ether (3 × 10 mL). The organic
layer was washed with water and brine, dried (Na2SO4),
filtered, and evaporated in vacuo to obtain a pale yellow liquid
that after radial chromatography (50% ethyl acetate-hexanes)
afforded a residue. This residue was dissolved in minimum
methanolic HCl and evaporated again in vacuo to yield a white
solid (24 mg, 71%): mp 180-192 °C (decomp); Rf 0.20 (50:50-
hexane:ethyl acetate, free base); 1H NMR (300 MHz, CD3OD)
δ 1.5241 (d, J ) 6.72 Hz, 3H), 3.0875 (dd, J ) 17.1, 10.86 Hz,
1H), 3.3547 (dd, J ) 16.62, 6.36 Hz, 2H), 4.25-4.40 (m, 1H),
7.5032 (d, J ) 7.35 Hz, 1 H), 7.5674 (t, J ) 7.71 Hz, 1 H),
7.8296 (t, J ) 7.8 Hz, 1 H), 7.9159 (d, J ) 7.83 Hz, 1 H), 9.0799
(s, 1 H); [R]D24 +2.58 (c ) 0.031, H2O).
1
122-123 °C; Rf 0.42 (50:50-hexane: ethyl acetate); H NMR
(80 MHz, CDCI3) δ 1.10 (d, J ) 7 Hz, 3H), 1.55 (s, 1H), 1.9 (s,
3H), 2.79 (dd, J ) 7, 3 Hz, 2H), 4.0-4.4 (m, 1H), 7.15-7.30
(m, 5H); IR (KBr) 3240, 3070, 2960, 1640, 1560, 1370, 1300,
1200, 1130, 770, 745, 700 cm-1; UV-vis (MeOH) λmax 206 (11
315) nm; MS (El) m/e 178 (M+ + 1, 1.7), 118 (18.6), 91(6.5), 86
(S)-3-Methyl-3,4-dihydroisoquinoline Hydrochloride.
This substance was prepared in 89.8% yield by the above
process, mp 180-192 °C (dec) following recrystallization from
MeOH/ether. TLC, IR, NMR, and MS agreed with those of the
prior specimen.
24
(17.4), 44 (100.00); [R]D +44.9° (c ) 0.2428, CHCl3);20,21
Analysis calcd for C11H14N0: C, 74.52; H, 8.52; N, 7.91;
found: C, 74.20; H, 8.80; N, 7.65.
(S)-(-)-Amphetamine Acetate. This compound was pre-
pared in 55% yield following the same procedure as for its
enantiomer. Following recrystallization from CH2Cl2/hexanes,
it gave mp 120-122 °C. TLC, IR, NMR, and MS agreed with
those of the prior specimen.
Ethyl N,N-Diethylaminomethylene Acetoacetate (19).
A solution of ethyl acetoacetate (1.3 g, 10 mmol), triethylortho-
formate (3 g, 20.3 mmol), and acetic anhydride (6.5 mL) were
heated under argon at 110 °C for 3 h after which the mixture
was evaporated under reduced pressure, azeotroped with small
amounts of toluene (5 mL × 3), and dried under high vacuum
for about 3 h. The crude product was redissolved in CH2CI2
(10 mL), and 2 mL diethylamine was added via syringe under
argon. After stirring the reaction mixture for 2 h at r.t., the
mixture was evaporated in vacuo and chromatographed (3:7
ethyl acetate-hexanes) to obtain 1.89 g (89%) of a dark green
oil which was used as such for condensation with 3-methyl-
3,4-dihydroisoquinoline: 1H NMR (300 MHz, CDCI3) δ 1.1 914
(br. s, 6H), 1.3219 (t, J ) 7.17 Hz, 3H), 2.2938 (s, 3H), 3.33-
3.41 (m, 4H), 4.2392 (q, J ) 7.17 Hz, 2H), 7.6386 (s, 1H); MS
(El) m/e 213 (M+, 18.0), 198 (26.8), 168 (40.8), 152 (91.1), 96
(67.4), 56 (81.0), 43 (100.0).
(R)-(-)-Amphetamine Hydrochloride (16). R-(+)-Am-
phetamine acetate (15, 2.860 g, 16.16 mmol) and 44% H2SO4
(16 mL) were refluxed for 40 h after which the mixture was
poured on ice and basified with NaOH pellets to pH 12-14.
The mixture was extracted into ether × 3 and washed with
5% HCl × 3. The aqueous layer was basified with 5% NaOH
solution and extracted into ether × 3. The ether solution was
concentrated, and dry HCl gas was bubbled through this
solution under an atmosphere of argon. Ether was evaporated
in vacuo to yield 2.330 g (84% crude) of a colorless solid: mp
1
153-154 °C (MeOH/ether); H NMR (80 MHz, CDCI3) δ 1.15
(d, J ) 7 Hz, 3H), 2.75 (d, J ) 8 Hz, 2H), 3.25-3.7 (m, 1H),
7.05-7.35 (m, 5H); [R]D -13.0° (c ) 0.0641, H2O).22,23
24
(R)-Ethyl 6-Methyl-1,6,7,1b-tetrahydro-2H-benzo[r]-
quinolizin-2-one-3-carboxylate (20). (R)-(+)-3-Methyldihy-
droisoquinoline hydrochloride (18, 1.129 g, 5.896 mmol), 19
(1.346 g, 6.319 mmol), and t-BuOH (40 mL) were refluxed
under argon for 48 h after which t-BuOH was evaporated off
under reduced pressure. The residue was taken up in CH2Cl2
and washed with 5% HCl (3 × 15 mL) and brine, dried (Na2-
SO4), filtered, and evaporated in vacuo to obtain a brown pasty
liquid which was purified by gravity column chromatography
eluting with 5% 2-propanol/ CH2CI2 to yield an off-white solid
(1.119 g, 66.6%) as a mixture of diastereomers (1.9:1). Part of
this product was recrystallized from CH2Cl2/ether/hexanes to
afford a colorless, crystalline solid: mp 128-129 °C; Rf ) 0.43
(S)-(-)-Amphetamine Hydrochloride. This was prepared
as above: mp 153-155 °C following recrystallization from
EtOH/ether. TLC, IR, NMR and MS agreed with those of the
prior specimen.
(R)-(+)-N-Formylamphetamine (17). (R)-(-)-Amphet-
amine hydrochloride (16, 3.069 g, 17.9 mmol) was dissolved
in EtOH (12 mL) under an argon atmosphere. Ethyl formate
(16.2 mL, 200.7 mmol) and Et3N (2.7 mL, 20.54 mmol) were
added dropwise in sequence, and the mixture was refluxed for
48 h. EtOH was evaporated in vacuo and residue taken up
into CH2CI2 (25 mL). The organic layer was washed with water
and 5% HCl (30 mL × 3), dried (Na2SO4), filtered, and
evaporated in vacuo to yield 1.806 g of crude product which
was chromatographed (gravity column) using 5% 2-propanol/
1
(5% i-PrOH-CH2Cl2); H NMR (300 MHz, CDCl3) δ 1.2 (d, J
) 3 Hz, 3H), 1.3 (t, J ) 6 Hz, 3H), 1.4 (d, J ) 3 Hz, 3H), 2.4-

1234 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 4
Fecik et al.
3.0 (m, 4H), 4.2 (q, J ) 6 Hz, 2H), 4.75-4.90 (m, 1 H), 7.0-
7.25 (m, 4H), 8.3 (s, 1H), 8.4 (s, 1H); lR (KBr) 3020, 2960, 1650,
1590, 1570 cm-1; UV-vis (MeOH) λmax 206 (32 419), 244
(37 834), 307 (37 596); Analysis calcd for C17H19N03: C, 71.54;
H, 6.72; N, 4.91; found: C, 71.24; H, 6.80; N, 5.10.
(S)-Ethyl 6-Methyl-1,6,7,1b-tetrahydro-2H-benzo[r]-
quinolizin-2-one-3-carboxylate. This compound was pre-
pared as white needles, mp 128-9 °C (from CH2Cl2), in 71.4%
yield as above. TLC, IR, NMR, and MS agreed with those of
the prior specimen.
1-(3,4-Difluorophenyl)-2-aminopropane (25). To a stirred
suspension of lithium aluminum hydride (3 equiv, 151 mmol,
5.72 g) in THF (130 mL) under N2
was added dropwise a
solution of nitropropene 24 (10.0 g, 50.2 mmol) in THF (100
mL) with cooling by an ice bath. After being refluxed 1 h, the
mixture was cooled to 0 °C and excess LAH was decomposed
by the sequential addition of H2O (20 mL), 20% NaOH (20 mL),
and H2O (40 mL), then the suspension was stirred until no
gray salts remained. The white salts were removed by filtra-
tion with suction, and the filter cake was washed with Et2O
(3×). The combined filtrate and washings were washed with
sat. NaCl (1 × 350 mL), dried (MgSO4), filtered, and evapo-
rated under reduced pressure to a mobile, yellow oil (7.78 g,
45.5 mmol, 91% yield): 1H NMR (300 MHz, CDCl3) δ 1.11 (d,
J ) 4 Hz, 3H), 1.26 (br s, 2H), 2.49 (dd, J ) 12, 8 Hz, 1H),
2.66 (dd, J ) 8, 4 Hz, 1H), 3.14 (m, 1H), 6.85-7.15 (m, 3H);
13C NMR (75.6 MHz, CDCl3) δ 23.7, 45.9, 48.7, 117.4 (d, J )
17 Hz), 118.2 (d, J ) 17 Hz), 125.37, 125.42, 125.44, 125.49;
(R)-Ethyl 6-Methyl-6,7-dihydro-2H-benzo[a]quinolizin-
2-one-3-carboxylate (21). (R)-20 (50 mg, 0.175 mmol), p-
chloranil (60 mg, 0.245 mmol) and dry benzene (3 mL) were
refluxed under argon for 12 h after which benzene was
removed under reduced pressure and the residue chromato-
graphed (gravity) using 5% 2-propanol-CH2Cl2 as the solvent
system to obtain 37 mg (81%) of a paste. This was scrubbed
with hexanes (3 mL) to obtain a gray powder (23 mg). A small
amount of this product was recrystallized from CH2Cl2-
hexanes to obtain colorless crystals: mp 125-127 °C; Rf ) 0.19
IR (neat) 3340, 3260, 2950, 2910, 1600, 1510, 1420, 1270 cm-1
;
MS (CI) m/z 172 (M+ + H), 141, 127, 44; calcd for C9H12F2N:
172.0913, found 172.0913.
1
(5% i-PrOH-CH2CI2); H NMR (80 MHz, CDCl3) δ 1.25 (t, J
1-(3,4-Difluorophenyl)-2-(N-formyl)aminopropane (26).
To amine 25 (2.00 g, 11.7 mmol) in Et2O (12 mL) cooled to 0
°C was slowly added acetic formic anhydride (1.6 mL) under
N2. After complete addition, the reaction was warmed to room
temperature and stirred for 18 h. The reaction mixture was
diluted with Et2O (25 mL) and washed with H2O (1 × 20 mL)
and sat. NaCl (1 × 20 mL). The ethereal layer was dried
(MgSO4), filtered, and evaporated under reduced pressure to
a red oil (2.32 g, 11.6 mmol, 99% yield): 1H NMR (300 MHz,
CDCl3) δ 1.15 (d, J ) 8 Hz), 1.29 (d, J ) 8 Hz, 3H for both),
2.64-2.88 (m, 2H), 4.39 (m, 1H), 5.85 (br s, 1H), 6.82-7.19
(m, 3H), 8.08 (s, 1H); 13C NMR (75.6 MHz, CDCl3) δ 20.3, 22.2,
41.9, 43.07, 45.5, 50.4, 117.6 (d, J ) 17 Hz), 118.5 (d, J ) 16
Hz), 125.55, 125.60, 125.62, 125.67, 161.0, 164.4; IR (neat)
3250, 3030, 2960, 2910, 2840, 1650, 1600, 1510, 1420, 1370,
1270 cm-1; MS (CI) m/z 200 (M+ + H), 154, 127, 72, 44; calcd
for C10H12F2NO: 200.0887, found 200.0895.
8,9-Difluoro-6,10b-dihydro-5-methyl-5H-oxazolo[2,3-a]-
isoquinoline-2,3-dione (27). To formamide 26 (28.19 g, 142
mmol) in CH2Cl2 (1.4 L) under N2 was added oxalyl chloride
(1.1 equiv, 156 mmol, 19.83 g, 13.9 mL) and the solution was
mechanically stirred at room temperature for 30 min. The
solution was cooled to -10 °C, and iron(III) chloride (1.2 equiv,
170 mmol, 27.64 g) was added. The mixture was slowly
brought to room temperature and stirred for 22.5 h. 2 M HCl
(1.4 L) was added to quench the reaction, and the biphasic
mixture was stirred at room temperature for 1 h. After the
layers were separated, the organic layer was washed with sat.
NaCl (1 × 700 mL), dried (Na2SO4), filtered, and evaporated
under reduced pressure to a dark red oil (30.46 g, 120 mmol,
85% yield): 1H NMR (300 MHz, CDCl3) δ 1.49 (d, J ) 4 Hz,
3H), 2.72 (dd, J ) 10, 4 Hz, 1H), 3.24 (dd, J ) 10, 4 Hz, 1H),
4.59 (m, 1H), 6.48 (s, 1H), 7.09 (dd, J ) 7, 4 Hz, 1H), 7.35 (dd,
J ) 6, 4 Hz, 1H); 13C NMR (75.6 MHz, CDCl3) δ 18.8, 34.1,
46.4, 81.0, 114.9 (d, J ) 19 Hz), 118.4 (d, J ) 18 Hz), 127.50,
127.56, 129.94, 130.00, 151.7, 158.6; IR (neat) 3050, 2980,
2930, 1815, 1730, 1510, 1415, 1325, 1310 cm-1; MS (CI) m/z
254 (M+ + H), 182, 166, 127; calcd for C12H10F2NO3: 254.0629,
found 254.0627.
) 6 Hz, 3H), 1.35 (d, J ) 7 Hz, 3H), 2.8 (dd, J ) 16, 3 Hz, 1H),
3.35 (dd, J ) 16, 5.5 Hz, 1H), 4.3 (q, J ) 6 Hz, 2H), 7.0 (s,
1H), 7.15-7.80 (m, 4H), 8.2 (s, 1H); IR (KBr) 2960, 1720, 1680,
1630 cm-1; UV-vis (MeOH) λmax 210 (19 348), 259 (30 225);
MS (El) m/e calcd for C17H17N03: 283.1208, found 283.1216;
284 (M+ + 1, 5.43), 253 (5.28), 239 (8.43), 211 (100.00), 167
24
(10.17); [R]D +2.53 (c 0.0103, CHCl3).
(S)-Ethyl 6-Methyl-6,7-dihydro-2H-benzo[a]quinolizin-
2-one-3-carboxylate. This substance was prepared in amor-
phous as above in 88% yield and was subjected directly to
hydrolysis without further purification. TLC, IR, NMR, and
MS agreed with those of the prior specimen.
(R)-6-Methyl-6,7-dihydro-2H-benzo[a]quinolizin-2-one-
3-carboxylic Acid (12). (R)-21 (358 mg, 1.265 mmol) was
dissolved in MeOH (3 mL) and 1 N NaOH (19 mL) was added
dropwise at r.t. After stirring the mixture for 8 h, it was
acidified to pH 1-2 with 1 N HCl when a colorless precipitate
fell out of solution. The solid was filtered under aspirator
vacuum and dried overnight on the high-vacuum pump to yield
262 mg (81%) of a colorless solid. Part of this product was
recrystallized from CH2Cl2 to obtain shiny, colorless crystals:
1
mp 250-251 °C; H NMR (300 MHz, CDCl3) δ 1.363 (d, J )
6.6 Hz, 3H), 2.955 (dd, J ) 16, 2.7 Hz, 1H), 3.492 (dd, J )
16.2, 5.4 Hz, 1H), 4.58-4.63 (m, 1H), 7.214 (s, 1H), 7.33-7.81
(m, 4H), 8.594 (s, 1H); IR (KBr) 3690-3090, 1700, 1640 cm-1
;
UV-vis (MeOH) λmax 258 (67 030), 210 (38 954) nm; MS (El)
m/e 256 (M+ + 1, 1.87), 211 (100.00), 167 (24.17), 139 (17.31),
24
115 (37.42), 89 (16.43), 63 (25.77), 53 (37.81); [R]D +189.4°
(c ) 0.0199, CHCl3); Analysis calcd for C15H13N03: C, 70.56;
24
H, 5.14; N, 5.49; found: 0, 70.60; H, 5.28; N, 5.61; [R]D
-185.1° (c ) 0.02035, CHCl3) for the S-isomer (prepared as
above in 21% yield following extensive recrystallizations from
CH2Cl2/hexane, mp 251-252 °C. TLC, IR, NMR, and MS
agreed with those of the prior specimen).
1-(3,4-Difluorophenyl)-2-nitropropene (24). A solution
of 3,4-difluorobenzaldehyde (23, 14.94 g, 105 mmol) and
ammonium acetate (67.5 mmol, 5.20 g) in nitroethane (250
mL) was refluxed 3 h. After removal of the solvent under
reduced pressure, the yellow residue was partitioned between
CH2Cl2 (300 mL) and H2O (300 mL), and the layers were
separated. The aqueous layer was extracted with CH2Cl2 (2 ×
200 mL), and the combined organic layers were washed with
sat. NaCl (1 × 200 mL), dried (Na2SO4), filtered, and evapo-
rated to produce a yellow oil (19.62 g, 98.5 mmol, 94% yield).
Upon standing, the oil solidified to bright yellow needles that
were recrystallized from EtOH to a constant melting point:
mp 46-48 °C; 1H NMR (300 MHz, CDCl3) δ 2.48 (s, 3H), 7.14-
7.33 (m, 3H), 7.79 (s, 1H); 13C NMR (75.6 MHz, CDCl3) δ 14.3,
71.4, 118.5 (d, J ) 18 Hz), 119.1 (d, J ) 18 Hz), 127.0, 127.0,
127.1, 131.6; IR (neat) 3050, 1650, 1590, 1510, 1410, 1315,
1290, 1270, 1110 cm-1; MS (CI) m/z 200 (M+ + H), 184, 168,
151, 141, 133; calcd for C9H9F2NO2: 200.0523, found 200.0518.
6,7-Difluoro-3,4-dihydro-3-methylisoquinoline (28). Ox-
azolidinedione 27 (30.46 g, 120 mmol) in concentrated H2SO4-
MeOH (1:19, 1.2 L) was refluxed for 20h. After cooling, the
solvent was evaporated under reduced pressure. The dark red
residue was partitioned between H2O (500 mL) and EtOAc
(500 mL). The organic layer was washed with 2 M HCl (2 ×
250 mL). The combined aqueous and acidic washings were
made basic with NH4OH and then extracted into CH2Cl2 (3 ×
250 mL). The combined organic extracts were washed with
sat. NaCl (1 × 500 mL), dried (Na2SO4), filtered, and evapo-
rated under reduced pressure to a dark yellow oil (15.79 g,
87.2 mmol, 73% yield): 1H NMR (300 MHz, CDCl3) δ 1.49 (d,
J ) 4 Hz, 3H), 2.48 (dd, J ) 10, 8 Hz, 1H), 2.75 (dd, J ) 10,
4 Hz, 1H), 3.70 (m, 1H), 6.96 (dd, J ) 6, 4 Hz, 1H), 7.12 (dd,

Chiral DNA Gyrase Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 4 1235
J ) 8, 6 Hz, 1H), 8.20 (d, J ) 2 Hz, 1H); 13C NMR (75.6 MHz,
CDCl3) δ 18.8, 34.1, 81.0, 114.9 (d, J ) 19 Hz), 118.4 (d, J )
18 Hz), 127.50, 127.56, 129.94, 130.00, 158.6; IR (neat) 3050,
2980, 2930, 1815, 1730, 1510, 1415, 1325, 1310 cm-1; MS (CI)
m/z 182 (M+ + H); calcd for C10H10F2N: 182.0781, found
182.0756.
with suction and dried under vacuum pressure overnight to a
light brown powder (240 mg, 0.83 mmol, 69% yield): 1H NMR
(300 MHz, TFA-d) δ 1.54 (d, J ) 4 Hz, 3H), 3.15 (br d, J ) 10
Hz, 1H), 3.65 (br d, J ) 10 Hz, 1H), 5.17 (m, 1H), 7.34 (br t, J
) 6 Hz, 1H), 7.78 (s, 1H), 7.88 (dd, J ) 4, 2 Hz, 1H), 9.24 (s,
1H); 13C NMR (125.8 MHz, TFA-d) δ 19.8, 34.1, 64.1, 112.1,
114.3, 114.6, 116.6, 118.6 (d, J ) 21.9 Hz), 118.9, 121.2 (d, J
) 19.0 Hz), 150.2, 152.0 (d, J ) 48.7 Hz), 152.8 (d, J ) 48.7
Hz), 170.7, 174.7; IR (KBr) 3390, 3000, 2940, 1685, 1610, 1490,
1410, 1290, 1250, 1190, 1140, 860 cm-1; MS (CI) m/z 292 (M+
+ H), 247; calcd for C15H12NO3F2: 292.0769, found 292.0785.
10-Fluoro-6-methyl-6,7,-dihydro-9-piperazinyl-2H-ben-
zo[a]quinolizin-2-one-3-carboxylic Acid (35). A mixture
of difluoride 34 (50 mg, 0.17 mmol), piperazine (1.5 equiv, 0.26
mmol, 22 mg), triethylamine (3 equiv, 0.51 mmol, 52 mg), and
acetonitrile (1.5 mL) was refluxed for 29.5 h. After cooling,
the volatiles were removed under reduced pressure. The
residue was triturated with Et2O and filtered with suction to
provide a dark brown solid (21 mg, 0.06 mmol, 35% yield): mp
216-217 °C (dec); 1H NMR (300 MHz, DMSO-d6) δ 1.12 (d, J
) 4 Hz, 3H), 2.74-2.88 (m, 9H), 3.01-3.10 (m, 1H), 4.39 (m,
1H), 6.93 (d, J ) 5 Hz, 1H), 7.26 (s, 1H), 7.86 (d, J ) 9 Hz,
1H), 8.76 (s, 1H); 13C NMR (125.8 MHz, TFA-d) δ 20.0, 34.4,
43.6, 46.7, 48.9, 49.0, 64.2, 113.0, 121.4, 121.9, 133.6, 145.1,
150.0, 153.4, 155.8, 157.8, 166.6, 171.0, 174.3; IR (KBr) 3520,
3440, 3370, 1620, 1600 cm-1; MS (CI) m/z 358 (M+ + H), 314,
292, 129, 128, 115, 87; calcd for C19H21FN3O3: 358.1541, found
358.1567.
Ethyl 2-Ethoxymethylene-3-trimethylsiloxy-3-buten-
oate (31). Fused zinc chloride (260 mg) was added to triethyl-
amine (2.2 equiv, 139 mmol, 14.10 g, 19.4 mL) and stirred for
1 h at room temperature under N2 until the salt was suspended
in the amine. To this was added ethyl (ethoxymethylene)-
acetoacetate (30, 11.79 g, 63.3 mmol) in benzene (14.2 mL)
and then chlorotrimethylsilane (2.0 equiv, 127 mmol, 13.80 g,
16.1 mL). After 30 min, the temperature was raised to 40 °C
and the reaction was stirred for 19.5 h. The mixture was
filtered with suction and then run through a plug of Celite
545 with the aid of Et2O. All remaining particulate matter was
removed by two gravity filtrations through three filter paper
circles, and the resulting solution was evaporated under
reduced pressure to a dark red oil (12.30 g, 47.6 mmol, 75%
yield): 1H NMR (300 MHz, CDCl3) δ 0.10 (s, 9H), 1.01-1.15
(m, 6H), 3.93-4.08 (m, 4H), 4.14 (d, J ) 2 Hz, 2H), 6.59 (s,
1H); 13C NMR (75.6 MHz, CDCl3) δ 0.0, 14.2, 15.2, 60.3, 70.3,
91.9, 97.6, 151.7, 153.1, 165.8; IR (neat) 2960, 2890, 1710, 1620,
1370, 1300, 1240, 1175, 1125, 1070, 1010, 840 cm-1; MS (CI)
m/z 259 (M+ + H), 207, 187, 145, 119; calcd for C12H23O4Si:
259.1368, found 259.1366.
Ethyl 9,10-Difluoro-6-methyl-1,6,7,11b-tetrahydro-2H-
benzo[a]quinolizin-2-one-3-carboxylate (32). To dihy-
droisoquinoline 28 (5.00 g, 27.6 mmol) in CH2Cl2 (80 mL) at
room temperature under N2 was added trifluoroacetic acid (1.0
equiv, 27.6 mmol, 3.15 g, 2.1 mL). After 5 min, boron trifluoride
diethyl etherate (1.0 equiv, 27.6 mmol, 3.92 g, 3.5 mL) was
added. After another 5 min, diene 31 (1.5 equiv, 41.4 mmol,
10.70 g) was added and the mixture was stirred at room
temperature for 2 h. The reaction was quenched by washing
with NaHCO3 (2 × 80 mL). The organic phase was separated,
dried (Na2SO4), filtered, and evaporated under reduced pres-
sure. The resulting dark oil was dissolved in MeOH (110 mL)
and stirred over anhyd Na2CO3 (2.76 g) at room temperature
for 1 h. The suspension was filtered and concentrated under
reduced pressure, and the residue was partitioned between
CH2Cl2 (100 mL) and H2O (100 mL). The organic phase was
dried (Na2SO4), filtered, and evaporated under reduced pres-
sure to a hard, dark brown residue (8.73 g, 27.2 mmol, 99%
yield) that was a complex mixture of diastereomers: 1H NMR
(300 MHz, CDCl3) δ 0.92-1.47 (m, 6H), 2.19-2.86 (m, 4H),
3.19-3.35 (m, 1H), 4.03-4.28 (m, 2H), 4.67-4.72 (m, 1H),
6.85-7.10 (m, 2H), 8.24 (s), 8.35 (s, 1H for both); MS (CI) m/z
322 (M+ + H), 276, 240, 194, 182, 127.
Ethyl9,10-Difluoro-6-methyl-6,7,-dihydro-2H-benzo[a]-
quinolizin-2-one-3-carboxylate (33). Dihydropyridone 32
(0.15 g, 0.47 mmol) and tetrachloro-1,4-benzoquinone (1.5
equiv, 0.71 mmol, 170 mg) in THF (2 mL) was refluxed for 23
h. The solvent was evaporated under reduced pressure, and
the crude product was purified by column chromatography (10:
1, MeOH-CHCl3) to give a dark brown residue (70 mg, 0.22
mmol, 47% yield): mp 134-135 °C (dec); 1H NMR (300 MHz,
CDCl3) δ 1.24-1.42 (m, overlapping d, 1.30, J ) 4 Hz, t, 1.36,
J ) 4 Hz, 6H), 2.82 (dd, J ) 1, 10 Hz), 1H), 3.37 (dd, J ) 2, 6
Hz, 1H), 4.29-4.50 (m, overlapping q, 4.29, J ) 4 Hz, m, 4.43,
3H), 6.86 (s, 1H), 7.09 (dd, J ) 5, 1 Hz, 1H), 7.52 (dd, J ) 8,
2 Hz, 1H), 8.24 (s, 1H); 13C NMR (75.6 MHz, CDCl3) δ 14.4,
19.9, 33.6, 57.0, 61.1, 114.9 (d, J ) 20 Hz), 118.1 (d, J ) 26
Hz), 123.7, 128.9, 141.5, 146.0, 148.0, 150.1, 151.2, 153.5,
164.6, 175.8; IR (KBr) 2970, 1720, 1625, 1500 cm-1; MS (CI)
m/z 320 (M+ + H), 247, 292; calcd for C17H16NO3F2: 320.1070,
found 320.1098.
10-Fluoro-6-methyl-6,7,-dihydro-9-(1,1-dimethylhydrazi-
nyl)-2H-benzo[a]quinolizin-2-one-3-carboxylic Acid (36).
A mixture of difluoride 34 (50 mg, 0.17 mmol) in 48%
tetrafluoroboric acid (1 mL) was heated at 90-100° for 19 h.
After being cooled to room temperature, the suspension was
poured into ice-water (5 mL) and filtered with suction. The
solid was washed with water and dried under vacuum pressure
to give a dark brown solid (55 mg). A mixture of this
intermediate boron difluoride chelate (55 mg, 0.16 mmol), 1,1-
dimethylhydrazine (1.5 equiv, 0.24 mmol, 25 mg, 19 µL),
triethylamine (3 equiv, 0.49 mmol, 50 mg, 69 µL), and
acetonitrile (1 mL) was refluxed for 24 h. After cooling, the
volatiles were removed under reduced pressure. The residue
was triturated with Et2O and filtered with suction to provide
a light brown solid (34 mg, 0.10 mmol, 64% yield): 1H NMR
(300 MHz, TFA-d) δ 1.70 (d, J ) 4 Hz, 3H), 3.28-3.90 (m,
8H), 5.33 (m, 1H), 7.43-7.56 (m, 1H), 7.93-8.38 (m, 3H); IR
(KBr) 3450 (br), 3400, 1695, 1625, 1605, 1300 cm-1; MS (CI)
m/z 332 (M+ + H), 329, 307, 289, 254, 232, 176, 154, 137; calcd
for C17H19N3O3F: 332.3487, found 332.341.
Acknowledgment. The authors gratefully acknowl-
edge partial financial support of this work by grants
from the NIH-AID Institute, USA, the Kansas Health
Foundation for a stipend to R.A.F., and the American
Foundation for Pharmaceutical Education for a stipend
to P.D.. We also thank Drs. Qun Li, Daniel T. W. Shu,
Jacob Plattner, and their associates at Abbott Labora-
tories for some of the microbiological evaluations.
References
(1) Gerster, J. R., SR; Pecore, S. E.; Winandy, R. M.; Stern, R. M.;
Landmesser, J. E.; Olsen, R. A.; Gleason, W. B. Synthesis,
absolute configuration and antibacterial activity activity of 6,7-
dihydro-5,8-dimethyl-9-fluoro-1-oxo-1H,5H-benzo[ij]quinolizine-
2-carboxylic acid. J. Med. Chem. 1987, 30, 839-843.
(2) Ishikawa, H.; Tabusa, F.; Miyamoto, H.; Kano, M.; Ueda, H.; et
al. Studies on antibacterial agents. Synthesis of substituted
6,I.;7-dihydro-1-oxo-1H,5H-benzo[i,j]quinolizine-2-carboxylic
acids. Chem. Pharm. Bull. (Tokyo) 1989, 37, 2103-2108.
(3) Morita, S.; Otsubo, K.; Matsubara, J.; Ohtani, T.; Kawano, Y.;
et al. Synthesis of possible metabolites of 1-cyclopropyl-1,4-
dihydro-6-fluoro-5-methyl-7-(3-methyl-1-piperazinyl)- 4-oxo-3-
quinolinecarboxylic acid (Grepafloxacin, OPC-17116). Chem.
Pharm. Bull. (Tokyo) 1995, 43, 2246-2252.
9,10-Difluoro-6-methyl-6,7,-dihydro-2H-benzo[a]quin-
olizin-2-one-3-carboxylic Acid (34). Ester 33 (380 mg, 1.2
mmol) was dissolved in MeOH (2.8 mL), and 1 M NaOH (18.5
mL) was added dropwise at room temperature and stirred for
5 h. The reaction was acidified to pH 1-2 with 1 M HCl, and
a solid precipitated from the solution. The solid was filtered
(4) Atarashi, S.; Yokohama, S.; Yamazaki, K.; Sakano, K.; Imamura,
M.; et al. Synthesis and antibacterial activities of optically active
ofloxacin and its fluoromethyl derivative. Chem. Pharm. Bull.
(Tokyo) 1987, 35, 1896-1902.

1236 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 4
Fecik et al.
(5) Hayakawa, I.; Atarashi, S.; Yokohama, S.; Imamura, M.; Sakano,
K.; et al. Synthesis and antibacterial activities of optically active
ofloxacin. Antimicrob. Agents Chemother. 1986, 29, 163-164.
(6) Gerster, J. R., Sr.; Rustad, N. J.; Reiter, M. J.; Pecore, S. E.;
Winandy, R. M.; Landmesser, J. E. Quinolones; J. R. Prous
Science Publishers: Barcelona, 1989; p 85.
(15) Chu, D. T.; Fernandes, P. B.; Claiborne, A. K.; Pihuleac, E.;
Nordeen, C. W.; et al. Synthesis and structure-activity relation-
ships of novel arylfluoroquinolone antibacterial agents. J. Med.
Chem. 1985, 28, 1558-1564.
(16) Lien, E. J.; Miranda, J. F.; Ariens, E. J. Quantitative structure-
activity correlation of optical isomers: a molecular basis for
Pfeiffer’s rule. Mol. Pharmacol. 1976, 12, 598-604.
(17) Larsen, R. D.; Reamer, R. A.; Corley, E. G.; Davis, P.; Grabowski,
E. J. J.; Reider, P. J.; Shinkai, E. A modified Bischler-Napieralski
procedure for the synthesis of 3-aryl-3,4-dihydroisoquinolines.
J. Org. Chem. 1991, 56, 6034-6038.
(18) Danishefsky, S.; Langer, M. E.; Vogel, C. On the use of the imine-
diene cyclocondensation reaction in the synthesis of yohimbine
congeners. Tetrahedron Lett. 1985, 26, 5983-5986.
(19) Kisfaludy, L.; Mohacsi, T.; Lowe, M.; Drexler, F. Pentafluoro-
phenyl acetate: a new, highly selective acetylating agent. J. Org.
Chem. 1979, 44, 654-656.
(7) Anon. European Patent 0 368 410 A2, 1995.
(8) Coughlin, S. A.; Danz, D. W.; Robinson, R. G.; Klingbeil, K. M.;
Wentland, M. P.; et al. Mechanism of action and antitumor
activity of (S)-10-(2,6-dimethyl-4-pyridinyl)-9-fluoro-3-methyl-
7-oxo-2,3-dihydro-7H-pyridol[1,2,3-de]-[1,4]benzothiazine-6-car-
boxylic acid (WIN 58161). Biochem Pharmacol 1995, 50, 111-
122.
(9) Fukuoka, Y.; Ikeda, Y.; Yamashiro, Y.; Takahata, M.; Todo, Y.;
et al. In vitro and in vivo antibacterial activities of T-3761, a
new quinolone derivative. Antimicrob. Agents Chemother. 1993,
37, 384-392.
(10) Miyadera, A.; Satoh, K.; Imura, A. Efficient synthesis of a key
intermediate of DV-7751 via optical resolution or microbial
reduction. Chem. Pharm. Bull. (Tokyo) 2000, 48, 563-565.
(11) Morrissey, I.; Hoshino, K.; Sato, K.; Yoshida, A.; Hayakawa, I.;
et al. Mechanism of differential activities of ofloxacin enanti-
omers. Antimicrob. Agents Chemother. 1996, 40, 1775-1784.
(12) Sato, K.; Hoshino, K.; Tanaka, M.; Hayakawa, I.; Osada, Y.;
Nishino, T.; Mitsuhashi, S. Molecular Biology of DNA Topo-
isomerases and Its Application in Chemotherapy, Proc. Int.
Symp.; CRC: Boca Raton, FL, 1993; p 167.
(13) Georgopapadakou, N. H.; Dix, B. A.; Angehorn, P.; Wick, A.;
Olson, G. L. Monocyclic and tricyclic analogs of quinolones:
mechanism of action. Antimicrob. Agents Chemother. 1987, 31,
614-616.
(20) Leonard, N. J.; Adamcik, J. A.; Djerassi, C.; Halpern, O.
Transannular nitrogen-carbon interaction in cyclic amino-
ketones and optical rotatory dispersion. J. Am. Chem. Soc. 1958,
80, 4858-4862.
(21) Welsh, L. H.; Report on dextro racemic amphetamines. J. Assoc.
Off. Agr. Chem. 1953, 36, 714-722.
(22) von Braun, J.; Friemelt, E. Abbau optisch-aktiver Carbonsaeuren
mit Stickstoffswasserstoffsaeure und Schwefelsaeure. Ber. 1933,
66, 684-685.
(23) Liethe, W. Die Konfiguration der Ephedrin-basen. Ber. 1932,
65, 660-666.
(24) Cavallito, C. J.; Grey, A. P. US Patent 3,489,840, 1970.
(25) Frank, K. E.; Devasthale, P. V.; Gentry, E. J.; Ravikumar, V.
T.; Keschavarz- Shokri, A.; Mitscher, L. A.; Nilius, A.; Shen, L.
L.; Shawar, R.; Baker, W. R. Combinatorial Chem., High
Throughput Screening 1998, 1, 73-83.
(14) Georgopapadakou, N. H.; Dix, B. A.; Angehorn, P.; Wick, A.;
Olson, G. L. Program Abstr. 25th Intersci. Conf. Antimicrob.
Agents Chemotherapy; ASM: Washington, DC, 1985; pp Abstr.
129.
JM0401356