Photochemistry of some fluoroquinolones: Effect of pH and chloride ion

Article abstract of DOI:10.1039/a903389k

A 6-fluoroquinolone (norfloxacin) and the naphthyridine analogue enoxacin give the corresponding 6-hydroxy derivatives by irradiation in water at pH 7.2 and, with lower efficiency, at pH 4.5 and 10. At pH 1 no defluorination takes place and the piperazinyl side chain is degraded. The 6,8-difluoro derivative lomefloxacin is defluorinated selectively from position 8 over the entire pH range considered (pH 1 to 10). The intermediate cation in position 8 does not add water and rather undergoes insertion into the β-CH bond of the neighboring N-ethyl group. The cation adds chloride, however. The structure-photoreactivity relationship for fluoroquinolones and the relation with the known phototoxicity of these compounds are commented upon.

Full text of DOI:10.1039/a903389k

View Article Online / Journal Homepage / Table of Contents for this issue
Photochemistry of some fluoroquinolones: effect of pH and
chloride ion
Elisa Fasani, Michela Rampi and Angelo Albini
Department of Organic Chemistry, University of Pavia, v. Taramelli 10, I-27100 Pavia, Italy.
E-mail: albini@chifis.unipv.it
Received (in Cambridge, UK) 28th April 1999, Accepted 12th July 1999
A 6-fluoroquinolone (norfloxacin) and the naphthyridine analogue enoxacin give the corresponding 6-hydroxy
derivatives by irradiation in water at pH 7.2 and, with lower efficiency, at pH 4.5 and 10. At pH 1 no defluorination
takes place and the piperazinyl side chain is degraded. The 6,8-difluoro derivative lomefloxacin is defluorinated
selectively from position 8 over the entire pH range considered (pH 1 to 10). The intermediate cation in position 8
does not add water and rather undergoes insertion into the β-CH bond of the neighboring N-ethyl group. The cation
adds chloride, however. The structure–photoreactivity relationship for fluoroquinolones and the relation with the
known phototoxicity of these compounds are commented upon.
Fluorinated organic derivatives are used for applications as
varied as new materials or pharmaceuticals due to the fact
that fluorine substitution imparts favourable properties while
increasing stability, because of the chemical inertness of the
carbon–fluorine bond. As an example, several fluoroaromatic
or heteroaromatic compounds have been shown to be superior
to the non-fluorinated derivatives in their pharmacological
activity and are actually used as drugs. A typical case is that of
the quinolone antibacterials, where the first generation drugs
such as nalidixic acid has been supplanted by fluorine-contain-
ing derivatives, nowadays amongst the most heavily prescribed
antibacterials. Due to our interest in the photochemistry of
heterocycles, we were attracted to this field by the numerous
recent reports of the light-related adverse side-effects of such
drugs. These have been reported to have phototoxic1 and
photoallergenic effects,2 as well as, in some cases, photo-
mutagenic and phototumorigenic effects.3 This is a major limit-
ation on their use in therapy.
in the presence of chloride which allows a better understanding
of the mechanism.
Results
Three fluoroquinolones were selected for this study, viz. 1-ethyl-
6-fluoro-7-piperazinyl-4-oxoquinoline-3-carboxylic acid (1,
norfloxacin), which may be taken as the simplest example of
the commonly used drugs of this group,17 a naphthyridine
analogue, enoxacin (2), and lomefloxacin (3), which, apart from
a methyl group in the chain, differs in having an additional
fluorine in position 8. In view of the amphoteric nature of these
molecules, conditions were chosen as to have one of the ionic
forms predominant.
Two sets of experiments were carried out, in both cases using
argon-flushed aqueous solutions of the fluoroquinolones. In
the first series, aimed at assessing the relative reactivity of these
derivatives, dilute [(1–2) × 10Ϫ4 M] solutions were irradiated
and the conversion was limited to 30%, in order to minimize
secondary photolysis. In the latter series, larger amounts of
saturated solutions were irradiated with the aim of separating
and identifying the photoproducts.
The chemical mechanism underlying such effects has not
been clarified up to now. A possible mechanism involves oxygen
sensitization by these drugs,4,5 but according to a recent study
by Chignell et al. this is not the main one.6,7
Another possibility is that the drugs react photochemically
with some cell component. Evaluation of this path requires that
the photochemistry of these drugs is ascertained. Positive evi-
dence about the photoreactions occurring has been scarce up to
now. It has been suggested that ciprofloxacin undergoes photo-
dimerization8 and in the case of (Ϫ)-ofloxacin several products
arising from the degradation of the alkylamino side-chain have
been isolated and characterized.9 In 1997–1998, several groups
reported that defluorination is the main photochemical reaction
from some of these drugs. Thus, Chignell et al.10 and Monti
et al.11 showed that irradiation of some fluoroquinolones leads
to release of fluoride anion, and our group (for the cases of
lomefloxacin, enoxacin and norfloxacin)12 as well as Mori-
muras one (for orbifloxacin)13 actually isolated and identified
the products of photodefluorination in water.
Effect of pH
Irradiation of a neutral solution of norfloxacin (1) (5 × 10Ϫ4 M,
NaHCO3 added to bring the pH at 7.2) gave a single major
product. We had previously shown in preparative experi-
ments12c that this was the hydroxyquinolone 4 through charac-
terization of the corresponding N-ethoxycarbonyl derivative
(Scheme 1). The effect of pH on the photoreaction was now
explored. In every case, the decomposition quantum yield was
measured and the fluoride liberated was potentiometrically
determined directly from the photolyzed solution. At pH 7.2,
Φdec was 0.06 and the molar amount of fluoride formed was
identical ( 5%) to that of decomposed 1. With solutions of 1 at
pH 10 (by addition of NaOH) or 4.5 (by addition of either HCl
or HClO4) the reaction course remained the same, with the
same peak in the HPLC and release of fluoride equivalent to
the substrate decomposition, but the quantum yield was lower
(Table 1).
Fluoroquinolones are amphoteric substrates,14,15 thus the
species present in solution depend on the pH. Thus, the photo-
chemistry is expected to be pH-dependent and there is some
literature indication that both the photodecomposition11,16 and
the oxygen photosensitization efficiency6 of some of these
derivatives depend on the pH. However, there is no indication
of whether the reaction path may be different at different pH.
We now report a product and quantum yield study on the
photochemistry of some fluoroquinolones at different pH and
On the other hand, when a solution at pH 1 (by addition of
either HCl or HClO4) was irradiated, the fluoride liberated was
<10% of the substrate decomposed. A preparative experiment
under these conditions led to the isolation of two products.
These were recognized as the fluorine-containing amino-
J. Chem. Soc., Perkin Trans. 2, 1999, 1901–1907
1901

View Article Online
Table 1 Photochemical reactions of fluoroquinolones 1 and 2
No added salt
0.2 M chloride
a
Substrate
pH
10
7.2
4.5
1
Φdec
FϪ (%)
Product
Φdec
FϪ (%)
Product
1
1
1
1
3
3
3
3
0.005 (0.007)
0.06 (0.009)
0.012 (0.01)
0.002
0.065
0.55 (0.55)
0.33
100
100
100
<10
100
100
93
4
4
4
5
10
10
10
10
0.007
0.06
0.01
100
100
<10
4
4
5
0.002
10
7.2
4.5
1
0.5
0.23
0.009
100
78
93
11
11
11
0.022
74
a In parentheses is the value in the presence of NaClO4 to make the total anion concentration 0.2 M. The products formed remain the same.
Fig. 1 (a) Decomposition of lomefloxacin (3) (᭜) and formation of
products 11 (᭿) and 1 (᭹) by irradiation of an aqueous solution 0.2 M
in NaCl at pH 7.2. (b) The same in 0.1 M HCl.
Ϫ
consistently defluorination (Φdec ≅ ΦϪF ) over the whole range
explored (pH 10, 7.2, 4.5 and 1). Compound 10 remained the
main product provided that HClO4 was used for acidification
(HCl gave a different result, see below). The decomposition
quantum yield for 3 was significantly larger than in the case of
the monofluoro analogue 1 [Φdec(3) = 0.55 at pH 7.2, see Table 1]
and was again reduced both by increasing and decreasing the
pH.
Scheme 1
quinolone 5 (which was obtained by chloroform extraction of
the irradiated solution) and 6 (recrystallized from the residue
after distillation of the solvent, see Scheme 1).
We extended the preparative experiments to the naphthyr-
idine derivative enoxacin (2). This compound under neutral
conditions is similarly defluorinated and gives phenol 7.12c
When the irradiation was carried out at pH 1, fluorine-
containing products were obtained, viz. the aminoquinolones
8 and 9 (see Scheme 1).
In the case of the 6,8-difluoroquinolone lomefloxacin (3) we
had found that irradiation under neutral conditions led to
selective defluorination from position 8, and the product
resulted not from solvolysis but from cyclization onto the
N-ethyl chain to give the pyrroloquinolinone 10 (Scheme 2).12c
The medium dependence of this reaction was now explored.
It was found that with this substrate the photoreaction was
Effect of added salts
Carrying out the above reactions at the same pHs as above
while maintaining a constant ionic strength by adding sodium
perchlorate so as to have a 0.2 M total ion concentration did
not change the course of the reactions and caused minor
changes in the quantum yield (see Table 1).
Experiments were also carried out in the presence of 0.2 M
chloride. In this case, quinolone 1 gave exactly the same results
as in the absence of added salt (Table 1), while the chemistry of
3 changed. With this substrate the quantum yield of decom-
position decreased somewhat with respect to the perchlorate
solution (particularly at low pH, see Table 1) and a new peak
was detected by HPLC. Preparative experiments on a neutral
solution 0.1 M in NaCl showed in fact that the main product
after 35 min irradiation (ca. 70% conversion of 3) was not 10
but the 6-fluoro-8-chloroquinolone 11 (Scheme 2), which was
isolated as the N-ethoxycarbonyl methyl ester. However, pro-
longed irradiation up to complete conversion of the starting
substrate gave again compound 10 as practically the only prod-
uct. Likewise, irradiation (7 h) of a solution 0.1 M in HCl gave
chlorinated 11, while longer irradiation gave 10.
The progress of the reaction could be conveniently moni-
tored by HPLC in small scale experiments, showing that in the
presence of chloride compound 11 was practically the only
product at the beginning of the irradiation, while the tricyclic
derivative 10 increased at the expense of the former at a rate
which became significant after most of 3 had been converted
[see Fig. 1(a) and (b) for neutral and acidic solutions
respectively].
Scheme 2
1902
J. Chem. Soc., Perkin Trans. 2, 1999, 1901–1907

View Article Online
Scheme 3
(protonation of the piperazine moiety) at pH 4.5. A second
protonation has not been characterized for the present
quinolones or for the parent 7-amino-4-quinolone. However,
taking into account the protonation equilibria of 1-methyl-4-
quinolone (pK 2.34)18 and 7-aminoquinoline (first pK 6.65,
second pK Ϫ0.03)19,20 it is reasonable to conclude that the
dication QH32ϩ is the main species at pH 1 (protonation at the
7-amino group is shown in Scheme 3, other tautomers may
be present). As appears in Table 1, the efficiency of the photo-
reaction changes significantly over the pH range explored.
With substrate 1, photoinduced solvolysis is the main process
in the pH range 4.5–10, though with decreased efficiency under
both basic and acidic conditions with respect to neutral condi-
tions. Related photoinduced aromatic substitutions occur either
via an addition–elimination mechanism [eqn. (1)], as demon-
Fig. 2 Proportion of products 10 (᭜) and 11 (᭿) as obtained at ca.
25% conversion by irradiation of an aqueous solution of drug 3 at pH
7.2 containing NaCl and NaClO4 (total salt concentration 0.2 M).
In order to have a more precise indication of the effect of
chloride concentration on the competition between the two
processes from 3, a further series of experiments with this drug
was carried out maintaining a constant ionic strength. Thus
irradiations were performed on a neutral solution in the
presence of NaCl–NaClO4 in various ratios with a total salt
concentration of 0.2 M. The corresponding change in the ratio
of products 10 and 11 (measured at ca. 25% conversion) is
represented in Fig. 2.
ArX* ϩ NuϪ → ArXNuϪ → ArNu ϩ XϪ (1)
strated for several nitroanisoles,21 or via a SN1 mechanism, as
recently suggested for electron-donating substituted halo-
benzenes,22 where the internal charge transfer character of the
excited state causes incipient heterolytic labilization on the
carbon–halogen bond [eqn. (2)]. The latter mechanism seems
ArX* ᎐ Arδϩ–XδϪ →
᎐
᎐
Arϩ ϩ XϪ Arϩ ϩ NuϪ → ArNu (2)
Discussion
Reactions of the quinolone excited state: structure and medium
dependence
appropriate for the present substrates, due to the presence of
the electron-donating amino group, and better explains some
characteristics of the reaction (see below). Cation 12 is then the
intermediate in the substitution to give 4 (Scheme 3). The
decrease in the quantum yield of reaction by changing the pH
is roughly proportional to the decrease in the proportion of
zwitterion QH present. A simple hypothesis would thus be that
the photonucleophilic substitution is characteristic of this
species, rather than of the singly charged ions, i.e. that hetero-
lytic C–F bond cleavage is more efficient from the zwitterion
QH (where it gives a cation) than from the cation QHϩ (giving a
dication) or in the anion QϪ (giving a non-stabilized zwitterion).
As has been mentioned in the introduction, it has recently
become clear that defluorination is the main process upon
irradiation of some fluoroquinolones.12,13 Actually, for the
presently considered derivatives 1–3 under most conditions,
Ϫ
defluorination is the only process (Φdec = ΦϪF , see Table 1).
However, there are important differences in the chemical pro-
cesses occurring and in the medium dependence among the
drugs studied. Thus, substitution of a hydroxy group for the
fluorine atom takes place with quinolones 1 and 2, except at pH
1, where degradation of the alkylamino side-chain is the main
process. With compound 3, on the other hand, defluorination
(selectively from position 8) occurs over the whole pH range
considered, but leads to intramolecular alkylation rather than
substitution by an OH group.
2ϩ
At pH 1 the dication QH3 is present and since the second
protonation—contrary to the first one—involves a nitrogen
atom directly linked to the aromatic π system, the degree of
internal charge transfer occurring in the excited state is
decreased (see Scheme 3). As a result, with quinolones 1 and 2
defluorination becomes a minor path, while the main reaction
involves the piperazinyl chain. This is a quite inefficient process
and may be envisaged as a hydrolysis of the C–N bond favored
by the positive charge on the nitrogen in the dication and giving
intermediate 13 (Scheme 3), though there are certainly other
possibilities. An inefficient degradation of the N-alkyl chain has
often been observed with arylalkylamines, e.g. in the presence
As for the pH dependence, the conditions were chosen in
order to explore the behavior of the different ionic forms of
these substrates. The dissociation constants for protonation and
deprotonation of norfloxacin (1) are 6.22 and 8.51 respectively,
and the corresponding values for lomefloxacin (3) are 5.49 and
8.78.14 Thus at pH ca. 7 these molecules are present essentially
as the zwitterions QH (see Scheme 3). On the other hand there
is ≥90% of the anion QϪ at pH 10 and of the monocation QH2
ϩ
J. Chem. Soc., Perkin Trans. 2, 1999, 1901–1907
1903

View Article Online
of excited hydrogen and electron acceptors such as ketones23
and aromatics24 under various conditions. It is thus not
unexpected that this reaction occurs, with Φ ≅ 10Ϫ3, when the
unimolecular decomposition of the heterocyclic moiety is
inhibited. In this way, fluorine-containing aminoquinolones 6
and, by further hydrolysis, 5 are obtained from 1, and analo-
gously 8 and 9 are obtained from 2.
With the difluoro derivative 3 defluorination occurs select-
ively from position 8, is 5 to 10 times more efficient than with 1
or 2 and remains the dominant process over the whole pH range
explored (1 to 10), though again the efficiency drops when get-
ting away from neutrality. We previously suggested that the
increased photolability introduced by further fluorine substitu-
tion in position 8 is due to the increased stability of the aryl
cation in this case.12b As is shown in Scheme 4, the σ aryl cation
Fig. 3 Ratio of the amount of product 10 formed in the absence ([10]0)
and in the presence of NaCl as obtained by irradiation of an aqueous
solution of drug 3 at pH 7.2 containing NaCl and NaClO4 (total
concentration 0.2 M).
4-XC6H4ϩ ϩ CF3CH2OH →
4-XC6H4OCH2CF3 ϩ Hϩ (3)
4-XC6H4ϩ ϩ CF3CH2OH →
ϩ
ؒ
4-XC6H5 ϩ CF CHOH (4)
˙
3
The 4-amino derivatives (X = NEt2, morpholino) behave as
electrophiles and give the trifluoroethyl ethers (as one would
expect from the singlet state of the cation) while the 4-benzoyl
derivative abstracts hydrogen (as expected from the triplet
state).26 Thus, seemingly strictly related cations may show dif-
ferent reactivity. As suggested before,12 with the quinolones this
is rationalized by the contribution of the π mesomer 14Ј to the
cation from 3. In this structure the charge is on the nitrogen
and position 8 has a carbene character (possibly a triplet), in
line with the observed intramolecular insertion into a C–H
bond (presumably a two-step homolytic process) (Scheme 4).
On the contrary, cations in position 6, such as those formed
from 1 and 2 (see formula 12 in Scheme 3) do not admit such
mesomerism and the localized σ cations behave as electro-
philes adding water. The present study shows that, while
water is unable to trap cation 14, a charged nucleophile such
as chloride ion adds to it, and under these conditions the
overall result is again fluorine substitution [to give chlorin-
ated 11, compare paths (a) and (b) in Scheme 4]. On the
other hand, chloride up to 0.2 M does not interfere with the
formation of phenol 4 from 1, consistent with an unselective
reaction from the less stabilized cation 12, which simply adds
to the most abundant nucleophile, as happens with the parent
phenylium cation.27,28
The competition between intermolecular chloride addition
and intramolecular insertion has been studied at different
chloride concentrations at constant ionic strength (0.2 M total
for NaClO4 and NaCl). As Fig. 2 and Table 1 show, path (b)
becomes significant at [ClϪ] = 1 × 10Ϫ3 M and is >90% at
[ClϪ] = 0.15 M. Under neutral conditions and constant ionic
strength the overall quantum yield changes very little with
different ratios of NaClO4 :NaCl, while the percentage of 11
changes from 0 to >90%. This supports the above hypothesis
that the photochemical process follows an SN1-type mechan-
ism and the intermediate cation 14↔14Ј partitions between
two paths. If ka and kb are the rate constants of the two
reactions, the ratio between the amount of compound 10
formed in the absence ([10]0) and in the presence ([10]) of
chloride depends linearly on the anion concentration accord-
ing to eqn. (5).
Scheme 4
in position 8 (see structure 14) is stabilised by the contribution
of a mesomeric π cation where the aromaticity of the pyridone
moiety is conserved (14Ј),† whereas this is not the case for a
cation in position 6, because in that case the π mesomer con-
tributes to a lesser degree, as the pyridone aromaticity is lost
in that structure (compare mesomer 15Ј in Scheme 4). This
explains both the regioselectivity in the photodefluorination of
3 and the lower reactivity of 6-monofluoro derivatives such as 1
and 2. The extra reactivity of fluorine in position 8 is also in
accord with the present finding that reaction at that position in
3 is less pH-dependent than that at position 6 in either 1 or 2.
Chemistry of the photochemically formed aryl cation
As mentioned above, photoinduced heterolytic C–F bond
cleavage leads to a different end result with 1 and 2 on one hand
(nucleophilic substitution by water) and with 3 on the other
(insertion in a C–H bond). A related case of dual chemistry
has been previously observed by Schuster et al. with some
4-substituted phenyl cations (resulting from the photodecom-
position of the corresponding phenyldiazonium salts) in
trifluoroethanol [eqns. (3) and (4)].26
[10]0/[10] = 1 ϩ kb/ka[ClϪ]
(5)
This expectation is borne out by experiment, as shown in
† The stabilisation of the pyridone structure via neutral and zwitter-
ionic mesomers is well known (ref. 25).
Fig. 3. This gives kb/ka = 80 MϪ1. Since the rate constant
1904
J. Chem. Soc., Perkin Trans. 2, 1999, 1901–1907

View Article Online
of chloride addition obviously cannot overcome the diffusion
constant, the cation must have a lifetime τ = 1/ka > 2.5 × 10Ϫ7
s.‡
is different. The photoinduced cleavage appears to be a neces-
sary consequence of the combination of electron donating and
accepting substituents which makes the C–F bond labile in the
excited state, particularly when in position 8.
The 8-choroquinolone 11 is, as one would expect, itself
photoreactive via C–Cl fragmentation at position 8 and
yields again cation 14. Therefore, prolonged irradiation of 3
gives the tricyclic derivative 10 also in the presence of chloride
through two photochemical steps via 11 (see Fig. 1). Efficient
defluorination from position 8 and the two-fold electrophile–
carbene chemistry of the resulting cation is not exclusive
to lomefloxacin. In fact, Murimura found that the 1-cyclo-
propyl-7-piperidino-5,6,8-trifluoro derivative orbifloxacin like-
wise undergoes selective monodefluorination.5,29 The product
obtained results from the substitution of a hydrogen for a fluor-
ine atom in position 8 and the loss of the cyclopropyl group
from position 1 (Scheme 5). This is probably due to the fact that
The above structure–reactivity relationship is supported by
literature data showing that insertion of a further donating
group strongly limits photoreactivity and phototoxicity, pre-
sumably by affecting the internal charge transfer character of
the excited state. Thus, an 8-alkoxy substituent has been con-
sistently found to reduce both the photolability and the photo-
toxicity as observed with ofloxacin (Φdec 0.001 at neutral pH)12b
and other alkoxy derivatives.31 More generally, all electron-
donating groups, independent of their position, have the same
effect (e.g. a 5-amino-6-fluoro-7-alkylamino-8-methyl derivative
has been found to be rather stable).32 Noteworthy is the case of
sparfloxacin, a 5-amino-7-dialkylamino-6,8-difluoroquinolone,
where the additional amino group in position 5 prevents the
photolability that would be expected for the fluoro group in
position 8.33
Conclusion
Some 6-fluoro- and 6,8-difluoro-7-aminoquinolones have been
found to photodecompose, with a higher efficiency in the latter
case. The reaction occurring is defluorination, and product
analysis and medium dependence suggest that this takes place
through an SN1 mechanism via an aryl cation. Although a
detailed quantum yield and product analysis has been carried
out only in a few cases, the sparse literature data are compatible
with the hypothesis that this is a general reaction. The cation in
position 6 adds water, while the cation in position 8 instead
undergoes a carbene insertion into the neighboring N-alkyl
chain. Addition is obtained also in position 8 when a charged
nucleophile such as choride is present. These drugs absorb in
the UV-A region and decompose when exposed to solar or
room light. They are known to be phototoxic and the efficiency
of photodefluorination correlates with the extent of the photo-
biological effect. This suggests that the reported phototoxic
effect of these drugs may be related to covalent binding of the
photochemically generated aryl cation intermediate to some
biological substrate. From the point of view of drug design, it
should be taken into account that the efficiency of the photo-
decomposition is related to the internal charge transfer char-
acter of the excited state.
Scheme 5
the cation undergoes intramolecular H abstraction from the
N-cyclopropyl group, analogously to the attack at the N-ethyl
group observed with 3, with the difference that the labile cyclo-
propyl moiety is degraded in this case. However, in the presence
of chloride anion, the 8-chloro substituted product conserving
the N-cyclopropyl chain was obtained also in this case in a
growing proportion for [ClϪ] = 0.05 to 0.2 M,29 again analo-
gously to 3.
Experimental
Norfloxacin (1) and enoxacin (2) were purchased from Sigma
Chemicals Co. (Milan) and used without further purification.
Lomefloxacin hydrochloride, from the same supplier, was dis-
solved in water (0.02 M solution) and 1 M NaOH was added to
neutrality. The free base (3) was extracted with chloroform (mp
231–234 ЊC). All other chemicals and solvents were reagent
grade or better. The pH was measured by means of a glass
electrode.
Relation with phototoxicity
This chemical evidence should be compared with the results
from photobiological research. It has been known for some
time that the presence of a halogen (F or Cl) at position 8
strongly enhances the phototoxicity of quinolone drugs.30
Recently, Chignell showed that 6,8-difluoro derivatives such as
lomefloxacin and fleroxacin are 10-fold more efficient in gener-
ating single strand DNA breaks after UV-A irradiation than are
6-fluoro derivatives such as norfloxacin.7b The fact that the
photoinduced defluorination of the species also occurs in the
same order, something which is not the case for their ability to
activate oxygen,6,7 coupled with the formation of highly reactive
intermediates in the defluorination process, suggests that the
biological effect may be due to covalent binding to some cell
component via such an intermediate. The chemical studies
show that an aggressive aryl cation is formed both from
6-fluoro and 6,8-difluoro quinolones, suggesting that the mech-
anism of phototoxicity remains the same, though the efficiency
Small-scale irradiations
Small scale photochemical reactions for quantum yield
determination and for the determination of the medium effect
were carried out on 10 ml portions of aqueous solution of
the drugs (0.1 mM) in serum capped quartz tubes. These
were irradiated in a merry-go-round apparatus by means of two
15 W phosphor-coated lamps (center of emission 313 nm).
Deoxygenation of the solution was obtained by flushing for 1 h
with argon passed through an appropriate furnace for elimin-
ating traces of oxygen.
The light flux was measured by ferrioxalate actinometry.
The substrate decomposition was determined by HPLC
(Waters Model 501 apparatus) with optical detection (Waters
‡ Dr S. Monti (CNR, Bologna, Italy) communicated to us the fact that
experiments are under way in her laboratory aimed at establishing the
formation of transients.
J. Chem. Soc., Perkin Trans. 2, 1999, 1901–1907
1905

View Article Online
490E, λ 276 nm). An inverse-phase Merck Purosphere RP-18 e
(5 µm) column (3 × 125 mm) was used. The eluant was an
85:15 mixture of pH 3 buffer (prepared by adding phosphoric
acid to a 0.75% solution of triethylamine until pH 3 was
reached) and acetonitrile. A few microlitres of a solution of an
appropriate standard (3 for 1, ofloxacin for 2 and 3) were added
to the photolyte before analysis.
The fluoride concentration was measured by means of an
Orion SA520 potentiometer using a selective electrode (Orion
F-94-09) after addition of 2 ml of Orion TISAB III buffer
(ammonium acetate, ammonium chloride, 1,2-cyclohexylidene-
α,δ-dinitrilotetraacetic acid) to the 10 ml photolyte and dilution
to 22 ml with distilled water.
δ 14.5, 14.9, 39.9, 41.8, 47.0, 51.9, 60.9, 110.9, 114.6, 117.0 (d,
J 17 Hz), 143.8, 146.8 (d, J 255 Hz), 147.2, 150.1 (d, J 15 Hz),
157.3, 166.1, 173.7. Compound 8 has been previously reported
but not characterized.36
A solution of lomefloxacin (3, 70 mg, 1.4 × 10Ϫ4 M) in 1.4 l
of 0.2 M NaCl solution was irradiated as above for 35 min. The
solution was extracted with 3 × 500 ml 1% ethyl choroformate
in chloroform and the organic layer was dried, concentrated to
10 ml, treated with excess ethereal diazomethane and evapor-
ated. Chromatography as above gave 8-chloro-1,4-dihydro-
7-(4-ethoxycarbonyl-3-methylpiperazinyl)-1-ethyl-6-fluoro-
quinoline-3-carboxylic acid methyl ester (11Ј, 11 mg) (Found:
C, 55.3; H, 5.7; N, 9.0. Calc for C21H25N3FClO5: C, 55.57; H,
5.51; N, 9.26%). 1H NMR [(CD3)2CO] δ 1.3 (t, J 7 Hz, 3H), 1.4
(d, J 7 Hz, 3H), 1.5 (t, J 7 Hz, 3H), 3.15 and 3.55 (two m, 2H),
3.2 and 3.3 (two m, 2H), 3.45 and 4.0 (two m, 2H), 3.75 (s, 3H),
4.15 (m, 2H), 4.75 and 4.82 (two m, 2H), 7.95 (d, J 12 Hz, 1H),
8.55 (s, 1H). 13C NMR (CDCl3) δ 14.6, 15.6, 15.9, 39.1, 47.4,
51.0, 52.2, 52.6, 55.2, 61.4, 110.6, 113.2 (d, J 22), 128.4, 135.5,
142.5 (d, J 15 Hz), 150.9, 152.4, 155.4, 156.8 (d, J 250 Hz),
165.9, 172.2. m/z 453 (base peak, contains 1 Cl). The non-
functionalized product 11 is known.37
Preparative irradiations
A solution of norfloxacin (1, 450 mg, 1 × 10Ϫ3 M) in 1.4 l of
0.1 M HCl was flushed for 1 h with purified argon and then
irradiated in an immersion well apparatus by means of a Pyrex-
filtered 500 W medium pressure mercury arc (Helios Italquartz)
at 25 ЊC while maintaining a slow flux of argon. The course of
the reaction was monitored by HPLC (see above). After 115 h
the acidic solution was extracted with an equal volume of
chloroform to give the known amine 5 (50 mg).34 The aqueous
phase was evaporated and the residue recrystallized from
ethanol to give 7-(2-aminoethylamino)-1-ethyl-6-fluoro-1,4-
dihydro-4-oxoquinoline-3-carboxylic acid (6, 70 mg), previ-
ously reported35 but not spectroscopically characterized. 1H
NMR [(CD3)2SO, 80 ЊC, at room temp., broad lines] δ 1.4 (t, J 7
Hz, 3H), 3.1 (m, 2H), 3.7 (m, 2H), 4.55 (q, J 7 Hz, 2H), 6.7 (br s,
exch, 1H), 6.95 (d, J 7 Hz, 1H), 7.71 (d, J 12 Hz, 1H), 8.3 (br s,
exch, 2H), 8.3 (s, 1H), 14.0 (br s, exch, 1H). A solution of
enoxacin (2, 640 mg, 2 × 10Ϫ3 M) in 1.1 l of 0.1 M HCl was
irradiated under argon as above for 12 h. A precipitate formed
during the irradiation and was filtered to give 7-amino-1-ethyl-
6-fluoro-1,4-dihydro-4-oxo-1,8-naphthyridine-3-carboxylic
acid (9, 80 mg), previously not spectroscopically character-
ized.36 1H NMR [(CD3)2SO] δ 1.3 (t, J 7 Hz, 3H), 4.4 (q, J 7 Hz,
2H), 7.8 (br s, exch, 2H), 7.9 (d, J 7 Hz, 1H), 8.3 (s, 1H). 13C
NMR [(CD3)2SO] δ 19.4, 51.5, 112.1, 115.5, 120.9 (d, J 17 Hz),
149.8 (d, J 255 Hz), 150.9, 151.4, 157.1 (d, J 17), 170.8, 180.0.
The aqueous solution was neutralized and extracted by 2 × 300
ml 1% ethyl chloroformate in chloroform. The organic layer
was washed with aqueous NaHCO3 and water, dried and con-
centrated. Excess ethereal diazomethane was added and the
residue was chromatographed on a silica gel column (eluent,
CHCl3–MeOH mixtures) to give, besides functionalized
unreacted starting material, 7-(2-ethoxycarbonylaminoethyl)-
1-ethyl-6-fluoro-1,4-dihydro-4-oxo-1,8-naphthyridine-3-carb-
oxylic acid methyl ester (8Ј, 60 mg). Mp 148–150 ЊC (toluene)
(Found: C, 57.1; H, 5.8; N, 11.0. Calc for C18H22N3FO5: C,
Acknowledgements
Financial support of this work by Istituto Superiore di Sanità,
Rome, in the frame of the ‘Program on the physical properties
of drugs and their safe use’, is gratefully acknowledged.
References
1 (a) J. Ferguson, Photochem. Photobiol., 1995, 62, 954; (b) P. Ball
and G. Tillotson, Drug Saf., 1995, 13, 343; (c) P. S. Lietman, Drugs,
1995, 49, 794; (d) S. R. Norrby and P. S. Lietman, Drugs, 1993, 45
(Suppl. 3), 59.
2 (a) T. Horio, H. Miyauchi, Y. Aoki and M. Harada, J. Dermatol.
Sci., 1994, 7, 130 (b) Y. Tokura, T. Nishijma, H. Yagi, F. Furukawa
and M. Takigawa, Photochem. Photobiol., 1996, 64, 838.
3 (a) G. Klecak, F. Urbach and H. Urwyler, J. Photochem. Photobiol.
B: Biol., 1997, 37, 174; (b) M. Maekinen, P. D. Forbes and
F. Stenbaek, J. Photochem. Photobiol. B: Biol., 1997, 37, 182;
(c) F. Urbach, J. Photochem. Photobiol. B: Biol., 1997, 37, 169; (d)
H. J. Reavy, N. J. Traynor and N. K. Gibbs, Photochem. Photobiol.,
1997, 66, 368.
4 (a) D. G. Robertson, G. A. Epling, J. S. Klely, D. L. Bailey and
B. Song, Toxicol. Appl. Pharmacol., 1992, 111, 221; (b) N. Wagai and
K. Tawara, Free Radical Res. Commun., 1992, 17, 387; (c) N. Wagai
and K. Tawara, Arch. Toxicol., 1992, 66, 392; (d) N. Umezawa,
K. Arakane, A. Ryu, S. Mashiko, M. Hirobe and T. Nagano, Arch.
Biochem. Biophys., 1997, 342, 275; (e) K. Shimoda, M. Nomura and
M. Kato, Fundam. Appl. Toxicol., 1996, 31, 133.
5 T. Morimura, K. Kohno, Y. Nobuhara and H. Matsukura, Chem.
Pharm. Bull., 1997, 45, 373.
6 P. Bliski, L. J. Martinez, E. B. Koker and C. F. Chignell, Photochem.
Photobiol., 1996, 64, 496.
7 (a) L. J. Martinez, R. H. Sik and C. F. Chignell, Photochem.
Photobiol., 1998, 67, 399; (b) L. Martinez and C. F. Chignell,
J. Photochem. Photobiol. B: Biol., 1998, 45, 51.
8 E. V. Tiefenbacher, E. Haen, B. Przybilla and H. Kurz, J. Pharm.
Sci., 1994, 83, 463.
1
56.99; H, 5.80; N, 11.8%). H NMR (CDCl3) δ 1.3 (t, J 7 Hz,
3H), 1.5 (t, J 7 Hz, 3H), 3.5 (m, 2H), 3.7 (m, 2H), 3.9 (s, 3H), 4.1
(q, J 7, 2H), 4.3 (q, J 7 Hz, 2H), 5.85 (br s, exch, 1H), 6.5 (br s,
exch, 1H), 7.9 (d, J 12 Hz, 1H), 8.4 (s, 1H). 13C NMR (CDCl3)
9 Y. Yoshida, E. Sato and R. Moroi, Drug. Res., 1993, 43, 601.
10 L. Martinez, G. Li and C. F. Chignell, Photochem. Photobiol., 1997,
65, 599.
11 S. Sortino, G. De Guidi, S. Giuffrida, S. Monti and A. Velardita,
Photochem. Photobiol., 1998, 67, 167.
12 (a) E. Fasani, M. Mella, D. Caccia, S. Tassi, M. Fagnoni and
A. Albini, Chem. Commun. 1997, 1329; (b) E. Fasani, A. Profumo
and A. Albini, Photochem. Photobiol., 1998, 68, 666; (c) E. Fasani,
F. F. Barberis Negra, M. Mella, S. Monti and A. Albini, J. Org.
Chem., 1999, 64, 5388.
13 T. Morimura, Y. Nobuhara and H. Matsukura, Chem. Pharm. Bull.,
1997, 45, 373.
14 K. Takacs-Novak, B. Noszal, I. Hermecz, G. Kereszturi, B. Podanyi
and G. Szasz, J. Pharm. Sci., 1990, 79, 1023.
15 (a) D. L. Ross and C. M. Riley, Int. J. Pharm., 1992, 83, 267;
(b) D. L. Ross and C. M. Riley, Int. J. Pharm., 1990, 63, 237.
1906
J. Chem. Soc., Perkin Trans. 2, 1999, 1901–1907

View Article Online
16 T. Morimura, T. Ohno, H. Matsukara and Y. Nobuhara, Chem.
Pharm. Bull., 1995, 43, 1000.
17 H. C. Neu, in The 4-Quinolones, ed. G. C. Crumplin, Springer-
Verlag, London, 1990, p. 1.
29 T. Morimura, K. Kohno, Y. Nobuhara and H. Matsuhura, Chem.
Pharm.Bull., 1997, 45, 1828.
30 J. M. Domagala, J. Antimicrob. Chemother., 1994, 33, 685.
31 (a) T. Shibata, M. Nagasawa, N. Iwai, M. Miyazaki, Y. Kawamura
and T. Kodama, Jpn. J. Antibiot., 1995, 48, 861; (b) M. Matsumoto,
K. Kojima, H. Nagano, S. Matsubara and T. Yokota, Anti-
microb. Agents Chemother., 1992, 36, 1715; (c) K. Marutani,
M. Matsumoto, Y. Otabe, M. Nagamura, K. Tanaka, A. Miyoshi,
T. Hasagawa, H. Nagano, S. Matsubara, R. Kamide, T. Yokota,
F. Matsumoto and Y. Ueda, Antimicrob. Agents Chemother., 1993,
37, 2217; (d) J. E. Rosen, D. Chen, A. K. Prahalad, T. E. Spratt,
G. Schluter and G. M. Williams, Toxicol. Appl. Pharmacol., 1997,
145, 381; (e) K. Shimoda, K. Akahane, M. Nomura and M. Kato,
Arzneim. Forsch., 1996, 46, 625.
18 I. Tucker, J. Am. Chem. Soc., 1951, 73, 1923.
19 A. Albert, R. Goldacre and J. N. Phillips, J. Chem. Soc., 1948, 2240.
20 E. V. Brown and A. C. Pasz, J. Heterocycl. Chem., 1970, 7, 335.
21 (a) J. Cornelisse, in CRC Handbook of Organic Photochemistry and
Photobiology, ed. W. M. Horspool and P. S. Song, Horspool, CRC,
Boca Raton, 1995, p. 250; (b) J. Cornelisse and E. Havinga, Chem.
Rev., 1975, 75, 353; (c) K. Mutai, R. Nakagaki and H. Tukada, Bull.
Chem. Soc. Jpn., 1985, 58, 2066; (d) H. C. H. A. Van Riel, G. Lodder
and E. Havinga, J. Am. Chem. Soc., 1981, 103, 7257; (e) A. M. J. Van
Eijk, A. H. Huizer, C. A. G. O. Varma and J. Marquet, J. Am. Chem.
Soc., 1989, 111, 88; ( f ) A. M. J. Van Eijk, A. H. Huizer and C. A. G.
O. Varma, J. Photochem., 1985, 29, 415.
32 T. Yoshida, Y. Yamamoto, H. Orita, M. Kakiuchi, Y. Takahashi,
M. Itakura, N. Kado, S. Yasuda, H. Kato and Y. Ito, Chem. Pharm.
Bull., 1996, 44, 1376.
33 Y. Tokura, Y. Iwamoto, K. Mizutani and M. Tanigawa, Arch.
Dermatol. Res., 1996, 288, 45.
34 C. B. Ziegler, W. V. Curran, N. A. Kuck, S. M. Harris and Y.-i. Lin,
J. Heterocycl. Chem., 1989, 26, 1141.
35 H. Koga, A. Itoh, S. Murayama, S. Suzue and T. Irikura, J. Med.
Chem., 1980, 23, 1358.
36 J-i. Matsumoto, T. Miyamoto, A. Minamida, Y. Nishimura,
H. Egawa and H. Nishimura, J. Med. Chem., 1984, 27, 292.
37 J. M. Domagala, A. J. Bridges, T. P. Culberson, L. Gambino,
S. E. Hagen, G. Karrick, K. Porter, J. P. Sanchez and J. A. Sesnie,
J. Med. Chem., 1991, 34, 1142.
22 (a) G. Zhang and P. Wan, J. Chem. Soc., Chem. Commun., 1994, 19;
(b) N. C. Yang, A. Huang and D. H. Yang, J. Am. Chem. Soc., 1989,
111, 8069; (c) A. P. Durand, R. G. Brown, D. Worrall and
F. Wilkinson, J. Chem. Soc., Perkin Trans. 2, 1998, 365.
23 A. H. Parola and S. G. Cohen, J. Photochem., 1980, 12, 41.
24 (a) D. Döpp and J. Heufer, Tetrahedron Lett., 1982, 23, 1553;
(b) M. Takami, T. Matsuura and I. Saito, Tetrahedron Lett., 1974,
661.
25 (a) A. R. Katritzky and J. M. Lagowski, Adv. Hetercycl. Chem.,
1963, 1, 339; (b) J. Elguero, C. Marzin, A. R. Katritzky and P. Linda,
The Tautomerism of Heterocycles, Academic Press, New York, 1976,
p. 71; (c) G. G. Hall, A. Hardisson and L. M. Jackman, Tetrahedron,
1963, 19, (Suppl. 2), 101.
26 S. M. Gasper, C. Devadoss and G. B. Schuster, J. Am. Chem. Soc.,
1995, 117, 5206.
27 J. P. Lorand, Tetrahedron Lett., 1989, 30, 7337.
28 H. Zollinger, Angew. Chem., Int. Ed. Engl., 1978, 17, 141.
Paper 9/03389K
J. Chem. Soc., Perkin Trans. 2, 1999, 1901–1907
1907