Photochemistry of 1-cyclopropyl-6-fluoro-1,4-dihydro-4-oxo-7-(piperazin-1-yl)quinoline-3- carboxylic acid (=ciprofloxacin) in aqueous solutions

Article abstract of DOI:10.1002/1522-2675(20010919)84:9<2508::AID-HLCA2508>3.0.CO;2-Y

The 1-cyclopropyl-6-fluoro-1,4-dihydro-4-oxo-7-(piperazin-1-yl)quinoline-3- carboxylic acid (=ciprofloxacin; 1) undergoes low-efficiency (φ=0.07) substitution of the 6-fluoro by an OH group on irradiation in H₂O via the ππ* triplet (detected by

Full text of DOI:10.1002/1522-2675(20010919)84:9<2508::AID-HLCA2508>3.0.CO;2-Y

2508
Helvetica Chimica Acta ± Vol. 84 (2001)
Photochemistry of 1-Cyclopropyl-6-fluoro-1,4-dihydro-4-oxo-7-
(piperazin-1-yl)quinoline-3-carboxylic Acid (ˆCiprofloxacin) in Aqueous
Solutions
by Mariella Mella, Elisa Fasani, and Angelo Albini*
Department of Organic Chemistry, University of Pavia Via Taramelli 10, I-27100 Pavia
Â
Dedicated to Professor Andre M. Braun on occasion of his 60th birthday.
The 1-cyclopropyl-6-fluoro-1,4-dihydro-4-oxo-7-(piperazin-1-yl)quinoline-3-carboxylic acid (ˆciprofloxacin; 1)
undergoes low-efficiency (F ˆ 0.07) substitution of the 6-fluoro by an OH group on irradiation in H2O via the
pp* triplet (detected by flash photolysis, lmax 610 nm, t 1.5 ms). Decarboxylation is a minor process (ꢀ 5%). The
addition of sodium sulfite or phosphate changes the course of the reaction under neutral conditions. Reductive
defluorination is the main process in the first case, while defluorination is accompanied by degradation of the
piperazine moiety in the presence of phosphate. In both cases, the initial step is electron-transfer quenching of
1
the triplet (kq ˆ 2.3 ´ 108m 1 s and 2.2 ´ 107m 1 s 1, respectively). Oxoquinoline derivative 1 is much more
photostable under acidic conditions, and in this case the F-atom is conserved, and the piperazine group is
stepwise degraded (F ˆ 0.001).
1. Introduction. ± A large number of 1-alkyl-7-(dialkylamino)-6-fluoro-1,4-dihydro-
4-oxoquinolinecarboxylic acids with various substituents at the benzo moiety are used
in therapy (3rd-generation fluoroquinolone antibacterials; quinolone ˆ quinolinone).
Overall, these are considered well-tolerated drugs; however, one of the important
exceptions is phototoxicity. This adverse effect has been reported for all of these
derivatives, though to a different degree, and in fact, it is recommended that exposure
to UV light should be minimized during therapy [1]. It is unusual that a class of drugs
presents light-related adverse effects of such seriousness (genotoxicity and tumori-
genesis have been reported [2]) and generality, and this has stimulated an intense effort
towards the elucidation of the biological mechanism underlying to such effects.
Different hypotheses are being considered at the moment, involving some form of
oxygen sensitization, the generation of radicals, or the formation of a reactive
intermediate, such as a carbene or an ion, by photofragmentation of the drug [3].
In parallel with biologically oriented work, photochemical studies have also been
carried out and have significantly contributed to the rationalization of this effect. The
reported photochemistry is highly dependent on the structure and on conditions. The
main processes observed are degradation of the alkylamino side chain [4 ± 6] and
substitution of the ring F-atom [4][6][7]. The latter process is quite efficient with some
derivatives of the series. Processes observed only in some cases are decarboxylation
[5d][8] as well as loss of the 1-alkyl chain, when this is a cyclopropyl group [6a][7b].
Product studies have been supplemented by photophysical investigations, mainly by
laser flash photolysis [4d,e][8 ± 10], aimed to establishing the complex mechanism
involved in the above photoreactions.

Helvetica Chimica Acta ± Vol. 84 (2001)
2509
The 1-cyclopropyl-6-fluoro-1,4-dihydro-4-oxo-7-(piperazin-1-yl)quinoline-3-car-
boxylic acid (ˆciprofloxacin; 1) is the most frequently used antimicrobial of this family
and has been early reported to be photounstable [11]. Mass-spectrometry evidence has
suggested that a photodimer of unknown structure is formed [12], while further studies
have led to the isolation of products resulting from the degradation of the piperazinyl
side chain [13]. On the basis of our studies on the photochemistry of related molecules
[4][8], the fact that defluorination had not been detected was difficult to account for.
This fact as well as our continuing interest in the field and the importance of compound
1 as a drug encouraged us to carry out a detailed study of its photochemistry.
2. Results. ± 2.1. Product Studies. Product studies were carried out by irradiating Ar-
flushed 3 ´ 10 4 m aqueous solutions of oxoquinoline derivative 1 [14] by means of a
Pyrex-filtered medium-pressure Hg arc (Scheme 1). The course of the reaction was
monitored by HPLC (see Exper. Part), showing the formation of a main peak and
several minor components. Extensively photodecomposed (ca. 80%) solutions were
Scheme 1

2510
Helvetica Chimica Acta ± Vol. 84 (2001)
analysed. The main product 2 was obtained by reversed-phase chromatography after
extraction of residual starting material and minor products by stirring with ethyl
carbonochloridate in CHCl3. The skeleton and the substitution pattern of 2 was
unchanged as compared to 1, but the F-atom was lost. Analytical and spectroscopic
properties showed that an OH group was present at position 6 of the quinoline moiety
of 2. In a separate experiment, extraction of the irradiated aqueous solution with
CHCl3 and recrystallization of the residue from the organic phase gave the most
abundant product 3 among the minor components. Compound 3 was again an F-free
compound, with position 6 unsubstituted and a deep-seated degradation of the
piperazine moiety to a [2-(formylamino)ethyl]methylamino group. The initial pH of
the solution was 6.2, and it decreased to 4.5 during the irradiation. The photochemistry
was also checked after addition of 5 ´ 10 4 m NaHCO3, in order to maintain the pH at
7.2 during the course of the reaction, and the products formed and the rate of reaction
were practically the same as above.
The irradiation was then carried out in 0.1n HCl. The reaction was much slower
under this condition, and the product distribution was different, as clearly shown by the
HPLC trace. At ca. 50% conversion, the 7-[(2-aminoethyl)amino]-6-fluoroquinoline
derivative 4 (Scheme 1) was the main product. At longer irradiation times, product 4
was degraded in turn, and the 7-amino-6-fluoroquinoline derivative 5 accumulated.
Both products 4 and 5 have been previously detected by Torniainen et al. [13c,d]. These
were formed only to a very minor extent under neutral conditions.
Following the approach we used with related molecules, the photochemistry of 1
was explored in saline solutions at neutral pH. In phosphate buffer (pH 7.4, 0.01m total
phosphate), the reaction was slower than in neat water, and product 2 was not formed.
On the contrary, compound 3, obtained in a small amount by irradiation in neat H2O,
was now the main product. Several other products were formed, one of which was
isolated as the N,N'-bis(ethoxycarbonyl) derivative of product 61) (Scheme 1).
The photochemistry in sulfite buffer (pH 7.2, 0.01m total sulfite) was again different.
Functionalization and extraction allowed to obtain the main product, identified as the
F-free derivative 71), and one of the minor ones, the fluorinated quinoline derivative 81)
(Scheme 1), both of them isolated as the corresponding N-(ethoxycarbonyl) deriva-
tives.
2.2. Kinetic Measurements. The reaction quantum yield was measured by using
1 ´ 10 4 m solutions of 1 and limiting the conversion to ca. 25%. The extent of conversion
of 1 and the formation of the F anion were measured. The measurements were
extended to the further conditions used in the preparative experiments. The results (see
Table) show that addition of sulfite or phosphate as well as of acids strongly decreased
the value of F. Furthermore, liberated fluoride corresponded to ca. 95% of the
substrate consumed in neutral H2O and remained >65% in the presence of sulfite or
phosphate buffer, but dropped to 25% in 0.1n HCl.
The effect of the above salts was also evaluated by measuring the yields of the main
product in the presence of phosphate (compound 3) or of sulfite (compound 7) vs. the
yield of the main product in their absence (compound 2), under different conditions. As
1
)
In Scheme 1, the initially formed products 6 ± 8 are shown; the actually isolated corresponding derivatives
are described in the Exper. Part.

Helvetica Chimica Acta ± Vol. 84 (2001)
2511
Table. Reaction Quantum Yield of the Conversion of 1 and Quantity of Fluoride Anion Formed under Different
Reaction Conditions
Neat H2Ob)
Phosphate buffer (0.007m)
Sulfite buffer (0.002m)
HCl (0.1n)
Fr
0.07
95
0.02
82
0.03
65
0.001
25
a
% F
)
a) With respect to the number of mol-equiv. of consumed compound 1. b) No difference in neat H2O (pH 6.2) or
in the presence of NaHCO3 solution (pH 7.2) was observed.
shown in Fig. 1, the ratios [3]/[2] and [7]/[2] linearly depended on the concentration of
the added salt.
The steady-state measurements were complemented by some time-resolved
investigations. Laser flash photolysis of an aqueous solution of 1 evidenced a
conspicuous transient in the ms range at ca. 610 nm (Fig. 2). This was not observed in
O2-equilibrated solutions. The transient was apparent at pH 7.2 (with NaHCO3) and at
pH 6.2, but under acidic conditions, the intensity was strongly reduced with a very weak
signal at pH ꢀ 4 (Fig. 3), nor was any other transient observed in the range 360 ±
650 nm. In the same pH range, the fluorescence shifted to the red (lmax from 408 to
446 nm). Under neutral conditions, the transient was quenched by addition of both
phosphate and sulfite. The measured quenching constants by these salts were kq ˆ 2.2 ´
1
1
107 m 1 s and 2.3 ´ 108 m 1 s , respectively.
3. Discussion. ± 3.1. Preamble. The present investigation showed that, under acidic
conditions, the main photochemical paths for ciprofloxacin (1) was degradation of the
piperazinyl side chain to give products 4 and 5, previously identified by Tornianinen et
3
al. [13c,d]. This was a quite inefficient reaction (F ˆ 1 ´ 10 ). On the other hand,
defluorination was by far the main process under neutral conditions, and led to
Fig. 1. Product distribution a) [3] vs. [2] and b) [7] vs. [2] on irradiation of compound 1 as a function of the
concentration of phosphate and sulfite buffer, respectively

2512
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 2. End-of-pulse absorption upon flashing an aqueous solution of compound 1. Inset: Time profile at 590 nm.
Fig. 3. Intensity of the end-of-pulse 590-nm absorption upon flashing an aqueous solution of compound 1 at
different pH values
photoproducts of different structure depending on conditions, in particular in the
presence of sulfite and phosphate.
3.2. Triplet State. The mechanistic evidence obtained and the comparison with the
results previously reported for related fluoroquinolinones offered a rationale for the
complex photochemistry of 1 observed. A transient was well apparent in neutral
aqueous solution, and wavelength distribution (lmax 610 nm), time profile (t 1.5 ms, see
Fig. 2), and sensitivity to O2 were analogous to what has been observed with
structurally related norfloxacin (aza-substituted at position 8) and ofloxacin (with
an alkoxy group at position 8) [9][10]. In view of the similarity and of the evi-
dence obtained in those cases, the transient was confidently assigned to the T-T
absorption.

Helvetica Chimica Acta ± Vol. 84 (2001)
2513
3.3. Sn (Ar) Reaction. The close correspondence between the modifications in the
course of the photoreaction and the quenching of this transient (see below) allowed to
conclude that the triplet state was the reactive state of 1 in neutral water. Substitution
of an F-atom by an OH group at the ring occurred with low efficiency (F 0.07), finally
giving product 2. Analogously to what has been observed with norfloxacin [4c], and
more generally with other photoinduced nucleophilic aromatic substitutions [15], the
process reasonably occurred via an addition-elimination mechanism (Sn2Ar* mecha-
nism, see Scheme 2, Path a), indicating the much stronger electrophilicity of the triplet
with respect to the ground state.
Scheme 2
3.4. Reduction of the Triplet. Both T-T absorption and formation of hydroxyquino-
line derivative 2 from 1 were efficiently quenched both by sulfite and by phophate (see
below for a quantitative evaluation). This was again analogous to what has been
observed with norfloxacin, and we offer the same rationalization based on the
reduction of the triplet (Scheme 2, Path b, X ˆ SO23 or HPO42 ). The triplet-reduction
potential, Ered(T) ˆ Ered ‡ Eel(T), was ca. 1.5 V vs. NHE for compound 1 and related
fluoroquinolinones (Ered
ˆ
1.3 V, Eel(T) ca. 2.8 eV) [3b][9][16]. This made reduction
.
.
by sulfite (E(SO23 /SO32 ) ˆ 0.63 V, E(SO3 /HSO3 ) ˆ 0.84 V) [17] a markedly exo-
.
thermic process and reduction by phosphate (E(HPO4 /HPO24 ) not available, but
estimated to be significantly lower than 1.9 V) [9][18] probably close to thermoneutral.
Such an evaluation was in accord with the highly efficient quenching of 13* by sulfite
1
1
(2.3 ´ 108 m 1 s ) and the still remarkable quenching by phosphate (2.2 ´ 107 m 1 s ).
.
The radical anion 1 formed from the electron-transfer step underwent defluori-
nation to yield an aryl radical 9. In the presence of a good reducing agent such as sulfite,
the sequence was completed by a second electron-transfer step and protonation leading
to F-free quinoline derivative 7 (see Scheme 2). In fact, with related fluoroquinoli-
nones, we found such a reductive defluorination to occur both by irradiation in the
presence of sulfite and by cathodic reduction [19].
This reaction did not take place in the case of phosphate, where the main product, 3,
resulted from reductive ring defluorination coupled with oxidative degradation of the
piperazinyl group. Since no products resulting from only one of these processes was
isolated (in contrast to other conditions, see below), a stepwise mechanism leading to
the final product should be proposed. A speculative mechanism is depicted in

2514
Helvetica Chimica Acta ± Vol. 84 (2001)
.
.
Scheme 3, where the phosphate radical anion (X ˆ HPO4 ) co-formed in the electron-
transfer step (and known as H-abstracting species [20]) abstracted a H-atom from the
side chain in the cage, and the diradical underwent ring cleavage and H2O addition to
finally give the (formylamino)ethyl derivative 3 or, through a more deep-seated
degradation, diamino derivative 6, the major and the minor product, respectively,
isolated in phosphate buffer. Both an analogue of 3 (from a related 1,8-naphtyridinone,
enoxacin) and an analogue of 6 (from a 6,8-difluoroquinolinone, lomefloxacin) have
been previously isolated [4c].
Scheme 3
Apart from the details of the mechanistic sequence, the steady-state analysis in
Fig. 1 connects quenching of triplet 1 by the above anions and formation of products 7
and 3. According to the competition shown in Scheme 4, the quantum yield for the
formation of product 2 (F2) and that for the formation of either product 7 or 3 (FX) in
the presence of the anions (X) are given by Eqn. 1 and 2, respectively, and their ratio by
Eqn. 3.
F2 ˆ Fisc kr/(kd ‡ kr ‡ kx [X])
FX ˆ Fisc kx [X]/(kd ‡ kr ‡ kx [X])
Fx/F2 ˆ kx [X]/kr
(1)
(2)
(3)
Scheme 4

Helvetica Chimica Acta ± Vol. 84 (2001)
2515
The products ratios [3]/[2] and [7]/[2] depended linearly on [X] (Fig. 1). These
1
1
plots give kx/kr ˆ 3.5 ´ 102 m
for phosphate and 4 ´ 103 m
for sulfite. Previous
determinations by the singlet depletion method suggested that Fisc ꢁ 0.5 with related
fluoroquinolinones [7c][20], although experiments based on energy transfer to
acenaphthenone indicated a lower value (Fisc ˆ 0.33) [10]. Taking the 0.5 value and
considering that in the absence of added anions, F2 ˆ Fisc kr/(kd ‡ kr) ˆ 0.07 (Table),
the value kr/(kd ‡ kr) ˆ 0.12 was obtained. On the basis of the measured rate of triplet
decay (kd ‡ kr) ꢂ 7 ´ 105 s , the pseudounimolecular rate of reaction with H2O (as the
1
solvent) was kr ꢂ 9 ´ 104 s . With the above ratios kx/kr, this gives kx ꢂ 3.1 ´ 107 m 1 s
1
1
for phosphate and 3.6 ´ 108 m 1 s for sulfite. The reasonable accord with the direct
measurements of kx (see above) supports that products 3 and 7 arise via quenching of
the triplet by anion X.
1
Another point is worthy of mentioning with regard to this mechanism. Compound 3
is formed also in the absence of phosphate, albeit in a lower amount (ꢀ1/5th than with
0.002m phosphate). A possible explanation is that H2O itself is oxidized inefficiently, as
.
.
expected from the highly positive Ered (T), generating the radical anion 1 and the OH
radical. The latter species may abstract a H-atom in the same way as the phosphate
.
.
radical anion (see Scheme 3, X ˆ OH ).
3.5. Decarboxylation. Another minor product worthy of commenting is compound 8
formed from 1 in the presence of sulfite. This had undergone reductive decarbox-
ylation, not defluorination. With fluoroquinolinones, decarboxylation has been
reported only in the case of the 8-(alkylthio) derivative rufloxacin, for which
defluorination was not the main photoproces (<40%) in H2O [8] [21]. In the present
case, decarboxylated 8 was isolated under conditions in which the triplet of 1 was fully
quenched and gave the radical anion. Decarboxylation could not result from this path,
and any rate would not be favored under reductive conditions. A possibility was that
decarboxylation was a minor process from the singlet (which would not be quenched
since tS was 1 ns) and gave radical 10 (Scheme 5). Under reductive conditions this led
to 8, while in neat H2O, it might form dimers, that we did not isolate, but had been
suggested to be among the products by Tiefenbacher et al. on the basis of HPLC/MS
evidence [12]. This remained a minor process, however, since 95% of the overall
photochemistry in H2O involved defluorination, not decarboxylation.
Scheme 5
3.6. Photochemistry in Acids. Finally, a different photochemistry was observed
under acidic conditions. In this case, no reaction occurred at the heteroaromatic moiety
of 1 (the amount of F liberated was only 25% of the overall reaction, see Table) and
the piperazine moiety was stepwise oxidatively degraded to give products 4 and 5 in an
3
inefficient process (F ˆ 1 ´ 10 ). A related ± and likewise inefficient ± degradation has

2516
Helvetica Chimica Acta ± Vol. 84 (2001)
been observed with lomefloxacin in 0.1n HCl [4d]. At pH 1, 1 was present in the mono-
and dicationic form rather than in the zwitterionic form predominating under neutral
conditions (pKa(1) 6.09 [22], see Scheme 6). This affected the nature of the excited
state (compare the shift in the fluorescence spectrum), which lost the internal charge-
transfer character. Under acidic conditions, the triplet was either not formed or had a
fully different character, since the T-T absorption, which was apparent under neutral
conditions, was not detected (see Fig. 3). The same observation has been made for
.
enoxacin [7c]. H-Abstraction either by triplet 1 or by OH radicals formed by
.
.
photoinduced electron transfer (cf. Scheme 2, X ˆ OH ) presumably causes the
observed degradation of the piperazinyl group, analogously to many known oxidative
degradations of dialkylanilines by photochemical, electrochemical, or thermal initia-
tion [23].
Scheme 6
4. Conclusions. ± It seems worthwhile to point out two conclusions from this work.
The first one refers to the identification of the photochemically labile functions on
oxoquinoline derivative 1 and the individuation of a structure/photoreactivity relation-
ship, in view of the worldwide use of this molecule as a drug and of the concern about
the photostability and phototoxicity of fluoroquinolinones and drugs in general [24].
The above results confirm that, analogously to related quinolinones [4][6], and
differently from what appeared from partial examinations [12][13], the main photo-
reaction in neutral solution is defluorination, though this may occur with a different
mechanism and lead to different end products in the presence of some anions. Fluorine-
conserving degradation of the piperazine group is important only under strongly acidic
3
conditions, and then with a low quantum yield (1 ´ 10 ). The 1-cyclopropyl group is
unaffected (and indeed a 1-cyclopropyl [6a][7b] and, more generally, a 1-alkyl group
[4a,c] has been found to be modified or eliminated in the photoreaction only in
quinolinones bearing an extra F-atom at position 8, for which a different mechanism
applies), and decarboxylation is only a minor process.
The second one refers to the mechanism. Fluoroquinolinones, and indeed most
highly stabilized six-membered heteroaromatics show very little unimolecular photo-
reactivity (again, 8-fluoro derivatives are an exception [4c][9]). However, as indeed it
is intuitively expected, the pp* triplet of these heterocycles is a strong electrophile and,
since this is a relatively long-lived species, a Sn2Ar* substitution occurs with some
efficiency in H2O. Obviously the triplet is also a strong oxidant, and is not only
efficiently reduced by sulfite but also, though at a rate one order of magnitude slower,

Helvetica Chimica Acta ± Vol. 84 (2001)
2517
by phosphate (a reaction for which there is little precedent [4e][25]) and even ± quite
inefficiently ± by H2O.
Importantly, the present data were obtained in Ar-flushed solution. Since most of
the reactions proceed via the triplet, they would be at least in part quenched in air-
equilibrated solutions, where different processes involving O2 may play a role.
Furthermore, the efficiency and the course of the photochemical reaction change
completely in phophate buffer, routinely used in (photo)biological studies. These two
facts illustrate that standard conditions for the study of thermal stability of drugs can
not be applied directly to the study of photostability or phototoxicity.
The photochemistry of (halo)heteroaromatics (frequent among drugs) in H2O has
been explored only to a very limited extent up to now, but a nucleophilic substitution
such as that observed here is expected to be a quite common result, and the same holds
for the electron-transfer path with inorganic anions. Since an F-atom at position 6
appears to be a required feature for a broad-spectrum antibacterial activity in
quinolinones (and has led to the success of ciprofloxacin) [14a][26], photoreactivity
(and thus light-induced degradation and toxic side effects) is an unavoidable
characteristic of these otherwise well tolerated [1] drugs.
Experimental Part
1. General. Column chromatography (CC): Merck silica gel (0.04 ± 0.063 mm) and CHCl3/MeOH; for
reversed phase, Fluka C18 reversed-phase silica gel (0.04 ± 0.063 mm) and phosphate buffer (pH 3)/MeOH.
HPLC: Hypersil column (4.6 Â 250 mm) and phosphate buffer (pH 3)/MeCN. IR Spectra: Perkin-Elmer 881
instrument; in cm 1. NMR Spectra: Bruker 300 instrument, d in ppm rel. to SiMe4 (ˆ0 ppm) as an internal
reference, J in Hz; attributions supported by exchange with D2O, double irradiation, BB, and DEPT-135
experiments when appropriate. Elemental analyses: Carlo-Erba 1106 instrument.
2. Photochemical Reactions. A soln. of cyprofloxacin (1, 129 mg) in 1.3 l of bidistilled H2O (3 ´ 10 4 m) in the
presence of the appropriate buffer in an immersion well apparatus was flushed with Ar for 40 min and then
irradiated by means of a Pyrex-filtered medium-pressure Hg arc (Helios Italquartz, 500 W), while maintaining a
slow flux of Ar. The course of the reaction was monitored by HPLC and the irradiation continued until ca. 80%
conversion was reached.
After irradiation in 0.1n HCl followed by extraction with CHCl3, 7-[(2-aminoethyl)amino]-1-cyclopropyl-
6-fluoro-1,4-dihydro-4-oxoquinoline-3-carboxylic acid (4) and 1-cyclopropyl-6-fluoro-7-amino-1,4-dihydro-4-
oxoquinoline-3-carboxylic acid (5) were obtained; the properties of these compounds corresponded to those
reported by Torniainen et al. [13c,d].
3. Irradiation in Neat Water. The irradiated soln. was stirred in CHCl3 (650 ml) for 5 h. The aq. phase was
evaporated and the residue submitted to reversed-phase CC: 1-cyclopropyl-1,4-dihydro-6-hydroxy-4-oxo-7-
(piperazin-1-yl)quinoline-3-carboxylic acid (2). Colorless crystals. M.p. 292 ± 2938. 1H-NMR ((CD3)2SO): 1.15 ±
1.4 (m, 4 H); 3.2 ± 3.7 (m, 8 H); 3.8 (m, 1 H); 7.45 (s, H C(8)); 7.65 (s, H C(5)); 8.3 (br. s, NH); 8.55
(s, H C(2)); 9.5 (br., OH); 10.7 (s, OH). 13C-NMR ((CD3)2SO): 7.8 (CH2); 36.0 (CH); 43.0 (CH2); 46.2 (CH2);
105.9 (CH(5)); 106.5 (C(3)); 109.0 (CH(8)); 120.4 (C); 136.1 (C); 145.7 (C(7)); 146.2 (CH(2)); 149.3 (C(6));
166.8 (COOH), 176.4 (C(4)ˆO). Anal. calc. for C17H19N3O4: C 61.99, H 5.82, N 12.76; found: C 62.3, H 6.0,
N 12.5.
The org. phase was stirred in H2O (50 ml) for 2 h, dried, and evaporated. Recrystallization from MeOH
gave 1-cyclopropyl-7-{[2-(formylamino)ethyl]methylamino}-1,4-dihydro-4-oxoquinoline-3-carboxylic acid (3).
Colorless crystals. M.p. 248 ± 2498. 1H-NMR ((CD3)2SO, 608): 1.1 ± 1.4 (m, 4 H); 3.15 (s, 3 H); 3.35 (q, J ˆ 7,
2 H); 3.65 (t, J ˆ 7, 2 H); 3.78 (m, 1 H); 7.05 (dd, J ˆ 2, 9, H C(6)); 7.15 (d, J ˆ 2, H C(8)); 7.9 (br., NH); 8.05
(s, CHO); 8.12 (d, J ˆ 9, H C(5)); 8.58 (s, H C(2)); 15.5 (br., COOH). 13C-NMR ((CD3)2SO, 608): 11.7
(CH2); 38.6 (CH2); 39.6 (CH); 45.2 (Me); 54.6 (CH2); 100.3 (CH); 110.1 (C(3)); 116.7 (CH); 118.5 (C); 131.1
(CH); 147.4 (C); 151.9 (C(2)); 158.9 (C(7)); 166.1 (CHO); 170.7 (COOH); 180.6 (C(4)ˆO). Anal. calc. for
C17H19N3O4: C 61.99, H 5.82, N 12.76; found: C 61.8, H 5.8, N 12.7.

2518
Helvetica Chimica Acta ± Vol. 84 (2001)
4. Irradiation in Phosphate Buffer. The irradiated soln. was stirred in CHCl3 (650 ml). The org. phase was
dried and evaporated and the residue recrystallized as above to give 3.
The aq. phase was again extracted with 350 ml of CHCl3 containing 1% of ethyl carbonochloridate, the org.
layer washed with H2O, dried, and evaporated, and the residue taken up with CHCl3 (20 ml) and treated with
diazomethane/Et2O. The soln. was evaporated and the residue submitted to CC: methyl 1-cyclopropyl-7-
{(ethoxycarbonyl){2-[(ethoxycarbonyl)amino]ethyl}amino}-1,4-dihydro-4-oxoquinoline-3-carboxylate, the
N,N'-bis(ethoxycarbonyl) methyl ester derivative of 61). Impure oil. IR: 1725. 1H-NMR (CDCl3): 1.15
(m, 2 H); 1.2 (t, J ˆ 7, 3 H ) ; 1.3 (t, J ˆ 7, 3 H ) ; 1.4 ( m, 2 H); 3.48 (m, 2 H); 3.5 (m, 1 H); 3.9 (m, 2 H); 3.95
(s, 3 H); 4.1 (q, J ˆ 7, 2 H); 4.25 (q, J ˆ 7, 2 H); 5.05 (br. NH); 7.35 (dd, J ˆ 2, 9, H C(6)); 7.8 (d, J ˆ 2,
H
C(8)); 8.45 (d, J ˆ 9, H C(5)); 8.6 (s, H C(2)). 13C-NMR (CDCl): 8.1 (CH2); 14.4 (Me); 14.5 (Me); 34.4
(CH); 39.8 (CH2); 49.6 (CH2); 52.0 (Me); 60.7 (CH2); 62.2 (CH2); 110.7 (C(3)); 114.0 (CH(8)); 123.3 (CH(6));
126.3 (C(4a)); 128.5 (CH(5)); 140.9 (C(8a)); 145.7 (C(7)); 149.0 (CH(2)); 155.2 (CON); 166.2 (COO); 173.6
(C(4)ˆO).
5. Irradiation in Sulfite Buffer. The irradiated soln. was stirred in 350 ml of CHCl3 containing 1% of ethyl
carbonochloridate for 5 h. The org. layer washed with H2O (50 ml), washed, dried, and evaporated, and the
residue taken up with CHCl3 (20 ml) and treated with diazomethane/Et2O. The soln. was evaporated and the
residue separated by CC into the derivatives of 7 and 8.
Data of Methyl 1-Cyclopropyl-7-[4-(ethoxycarbonyl)piperazin-1-yl]-1,4-dihydro-4-oxoquinoline-3-carbox-
ylate, the N'-(ethoxycarbonyl) methyl ester derivative of 71). Colorless crystals. M.p. 193 ± 1948. IR: 1720.
1H-NMR (CDCl3): 1.15 (m, 2 H); 1.35 (m, 2 H); 1.3 (t, J ˆ 7, 3 H); 3.35 (m, 4 H); 3.4 (m, 1 H); 3.7 (m, 4 H); 3.9
(s, 3 H); 4.2 (q, J ˆ 7, 2 H); 7.05 (dd, J ˆ 2.5, 9, H C(6)); 7.15 (d, J ˆ 2.5, H C(8)); 8.35 (d, J ˆ 9, H C(5)); 8.55
(s, H C(2)). 13C-NMR (CDCl3): 8.0 (CH2); 14.5 (Me); 34.1 (CH); 43.1 (CH2); 47.6 (CH2); 51.5 (Me); 61.6
(CH2); 99.4 (CH(8)); 110.4 (C(3)); 113.9 (CH(6)); 120.9 (C(4a)); 128.8 (CH(5)); 142.1 (C(8a)); 148.6 (CH(2));
153.6 (C(7)); 155.3 (O C(ˆO)N); 166.6 (COO); 173.6 (C(4)ˆO). Anal. calc. for C21H25N3O5: C 63.14, H 6.31,
N 10.52; found: C 63.2, H 6.3, N 10.5.
Data of 1-Cyclopropyl-7-[4-(ethoxycarbonyl)piperazin-1-yl]-6-fluoro-quinolin-4-(1H)-one ( ˆ Ethyl 4-(1-
cyclopropyl-6-fluoro-1,4-dihydro-4-oxoquinolin-7-yl)piperazine-1-carboxylate), the N'-(ethoxycarbonyl) deriv-
ative of 81) (which in turn is known [27]). Impure oil. IR: 1720. 1H-NMR (CDCl3): 1.15 (m, 2 H); 1.3 (t, J ˆ 7,
3 H); 1.35 (m, 2 H); 3.25 (m, 4 H); 3.4 (m, 1 H); 3.7 (m, 4 H); 4.2 (q, J ˆ 7,2 H); 6.2 (d, J ˆ 8, H C(2)); 7.65
(d, J ˆ 8, H C(3)); 7.3 (d, J ˆ 4, H C(8)); 8.0 (d, J ˆ 12, H C(5). 13C-NMR (CDCl3): 8.1 (CH2); 14.6 (Me);
33.5 (CH); 43.4 (CH2); 49.9 (CH2); 61.6 (CH2); 104.5 (CH(8)); 109.3 (CH(3)); 112.3 (d, J(C,F) ˆ 22, CH(5));
121.5 (C(4a)); 138.9 (C(8a)); 141.2 (CH, CH(2)); 144.2 (d, J(C,F) ˆ 15, CH(7)); 152.7 (d, J(C,F) ˆ 250, CF(6));
155.3 (O C(ˆO)N); 176.8 (C(4)ˆO).
6. Small-Scale Photolyses. In quartz tubes, a 10-ml portion of the appropriate 1 ´ 10 4m aq. soln. of 1 was
flushed with Ar for 15 min and then capped. The soln. was irradiated in a merry-go-round apparatus fitted with
6 Â 15 W phosphorus-coated low-pressure lamps, centre of emission 310 nm. The product composition was
determined by HPLC with appropriate calibration curves (ofloxacin as the internal standard) and light flux by
ferrioxalate actinometry. The fluoride concentration was measured by means of an Orion-SA-520 potentiometer
with a selective electrode (Orion F-94-09), after addition to the photolysate (10 ml) of 2 ml of Orion-TISAB-III
buffer (ammonium acetate, ammonium chloride, 2,2',2'',2'''-[(cyclohexane-1,2-diyl)dinitrilo]tetrakis[acetic acid]
and dilution to 22 ml with bidistilled H2O.
7. Laser Flash Photolysis. Nanosecond flash photolysis experiments were performed by means of a
Nd :YAG JK laser (pulse 20 ns full width at half-maximum, l 266 nm) in the previously described experimental
setup [4c]. Ar-Flushed solns. were used with an absorbance of ca. 0.5 at 266 nm. The incident pulse energies were
< 4 mJ pulse 1. The spectra were reconstructed point-by-point from time profiles taken each 10 nm.
Á
Financial support of this work by the Istituto Superiore di Sanita, Rome within the ꢀProgram on the Physical
Properties of Drugs and Their Safe Useꢁ, is gratefully acknowledged.
REFERENCES
[1] R. Stahlmann, R. Lode, Drugs 1999, 58 (Suppl. 2), 37.
[2] H. J. Reavy, N. J. Traynor, N. K. Gibbs, Photochem. Photobiol. 1997, 66, 368; S. J. Bulera, J. C. Theiss, T. A.
Festerling, F. A. de la Iglesia, Toxicol. Appl. Pharmacol. 1999, 156, 222; M. Makinen, P. D. Forbes, F.
Stenback, J. Photochem. Photobiol. B: Biol. 1998, 37, 182.

Helvetica Chimica Acta ± Vol. 84 (2001)
2519
[3] a) L. Martinez, C. F. Chignell, J. Photochem. Photobiol. B: Biol. 1998, 45, 51; b) S. Sortino, G. Condorelli,
G. De Guidi, S. Giuffrida, Photochem. Photobiol. 1998, 68, 652; c) J. E. Rosen, D. Chen, A. K. Prahalad,
T. E. Spratt, G. Schluter, G. M. Williams, Toxicol. Appl. Pharmacol. 1997, 145, 381; d) N. Umezawa, K.
Arakane, A. Ryu, S. Mashiko, M. Hirobe, T. Nagano, Arch. Biochem. Biophys. 1997, 342, 275; e) G.
Condorelli, G. De Guidi, S. Giuffrida, P. Miano, S. Sortino, A. Velardita, Med. Biol. Environ. 1996, 24, 103.
[4] a) E. Fasani, M. Mella, D. Caccia, S. Tassi, M. Fagnoni, A. Albini, Chem. Commun. 1997, 1329; b) E. Fasani,
A. Profumo, A. Albini, Photochem. Photobiol. 1998, 68, 666; c) E. Fasani, F. F. Barberis Negra, M. Mella, S.
Monti, A. Albini, J. Org. Chem. 1999, 64, 5388; d) E. Fasani, M. Rampi, A. Albini, J. Chem. Soc., Perkin
Trans. 2 1999, 1901; e) E. Fasani, M. Mella, S. Monti, A. Albini, Eur. J. Org. Chem. 2001, 391.
[5] a) L. Elrod, C. L. Linton, B. P. Shelat, C. F. Wong, J. Chromatogr. 1990, 519, 125; b) Y. Yoshida, E. Sato, R.
Moroi, Arzneim.-Forsch. 1993, 43, 601; c) K. Demachi, H. Maeda, Y. Todo, K. Ojima, T. Kokei, H.
Imaizumi, M. Tai, Y. Nakajima, H. Narita, Yakugaku Zasshi 1995, 115, 716; d) N. Mizushima, F. Ohuchi, T.
Fujii, Jpn. J. Hosp. Pharm 1990, 16, 198.
[6] a) T. Morimura, Y. Nobuhara, H. Matsukura, Chem. Pharm. Bull. 1997, 45, 373; b) T. Morimura, T. Ohno,
H. Matsukara, Y. Nobuhara, Chem. Pharm. Bull. 1995, 43, 1000; c) T. Morimura, K. Kohno, Y. Nobuhara,
H. Matsuhura, Chem. Pharm. Bull. 1997, 45, 1828.
[7] a) L. J. Martinez, G. Li, C. F. Chignell, Photochem. Photobiol. 1997, 65, 599; b) M. Engler, G. Ruesing, F.
Soergel, U. Holzgrabe, Antimicrob. Agent Chemother. 1998, 42, 1151; c) S. Sortino, G. De Guidi, S.
Giuffrida, S. Monti, A. Velardita, Photochem. Photobiol. 1998, 67, 167.
[8] G. Condorelli, G. De Guidi, S. Giuffrida, S. Sortino, R. Chillemi, S. Sciuto, Photochem. Photobiol. 1999, 70,
280.
[9] S. Monti, S. Sortino, E. Fasani, A. Albini, Chem. ± Eur. J. 2001, 7, 2501.
[10] S. Navaratnam, J. Claridge, Photochem. Photobiol. 2000, 72, 283.
[11] J. Ferguson, G. Phillips, J. McEwan, T. Moreland, B. E. Johnson, Brit. J. Dermatol. 1988, 4, 119; M. E.
Hidalgo, C. Pesoa, E. Fernandez, A. M. Cardenas, J. Photochem. Photobiol. A: Chem. 1993, 73, 135.
[12] E. M. Tiefenbacher, E. Haen, B. Przybilla, H. Kurz, J. Pharm. Sci. 1994, 83, 463.
[13] a) S. Tammilehto, H. Salomies, K. Torniainen, J. Planar Chromatogr. 1994, 7, 368; b) K. Torniainen, S.
Tammilehto, V. Ulvi, Int. J. Pharm. 1996, 132, 53; c) K. Torniainen, J. Mattinen, C. P. Askolin, S.
Tammilehto, J. Pharm. Biomed. Anal. 1997, 15, 887; d) K. Torniainen, C. P. Askolin, J. Mattinen, J. Pharm.
Biomed. Anal. 1997, 16, 439.
[14] J. P. Sanchez, J. M. Domagala, S. E. Hagen, C. L. Heifetz, M. P. Hutt, J. B. Nichols, A. K. Trehan, J. Med.
Chem. 1988, 31, 983; A. M. Hay, S. Hobbs-Dewitt, A. A. MacDonald, R. Ramage, Synthesis 1999, 1979.
[15] J. Cornelisse, in ꢀCRC Handbook of Organic Photochemistry and Photobiologyꢁ, Eds. W. H. Horspool and
P. S. Song, CRC Press, Boca Raton, 1995, p. 2650; A. M. J. Van Eijk, A. M. Huizer, C. A. G. O. Varma, J.
Marquet, J. Am. Chem. Soc. 1988, 111, 88.
[16] P. Bilski, L. J. Martinez, E. B. Koker, C. F. Chignell, Photochem. Photobiol. 1996, 64, 496.
[17] R. E. Huie, P. Neta, J. Phys. Chem. 1984, 88, 5665.
[18] D. Stanbury, Adv. Inorg. Chem. 1989, 35, 69; P. Maruthamuthu, P. J. Neta, J. Phys. Chem. 1978, 82, 710.
[19] A. Profumo, E. Fasani, A. Albini, Heterocycles 1999, 51, 1499.
[20] S. Steenken, L. Goldbayarana, J. Am. Chem. Soc. 1998, 120, 3928; M. Nakashima, E. Hayon, J. Phys. Chem.
1970, 74, 3290.
[21] S. Sortino, G. Marconi, S. Giuffrida, G. De Guidi, S. Monti, Photochem. Photobiol. 1999, 70, 280.
[22] D. L. Ross, C. M. Riley, Int. J. Pharm. 1992, 83, 267.
[23] A. H. Parola, S. G. Cohen, J. Photochem. 1980, 12, 41; D. Doepp, J. Heufer, Tetrahedron Lett. 1982, 23,
1553; M. Takani, T. Matsuura, I. Saito, Tetrahedron Lett. 1974, 661; T. W. Bentley, D. J. Richards, M. G.
Hutchings, Tetrahedron Lett. 1986, 27, 5261; E. Fasani, T. Soldi, A. Albini, S. Pietra, Gazz. Chim. Ital. 1990,
120, 109.
[24] ꢀDrugs. Photochemistry and Photostabilityꢁ, Eds. A. Albini and E. Fasani, Royal Soc. Chem., Cambridge,
1998; ꢀPhotostability of Drugs and Drug Formulationsꢁ, Ed. H. Tùnnesen, Taylor & Francis, London, 1996.
[25] V. A. Kuzmin, A. K. Chibisov, Dokl. Akad. Nauk SSSR 1973, 212, 1146.
[26] T. D. Gootz, K. E. Brighty, Med. Res. Rev. 1996, 433; H. Koga, A. Itoh, S. Murayama, S. Suzue, T. Irikura, J.
Med. Chem. 1980, 23, 1358.
[27] M. Renmann, M. A. Eisenstat, J. D. Weaver, Tetrahedron Lett. 1994, 35, 8303.
Received May 23, 2001