Toxicology Mechanisms and Methods, 13: 221–226, 2003Copyright c© Taylor & Francis Inc.ISSN: 1537-6524 print / 1537-6516 onlineDOI: 10.1080/15376520390223480

Photo-Induced Interaction of Antibacterial QuinolonesWith Human Serum Albumin

Franklin Vargas and Carlos RivasCentro de Quımica, Instituto Venezolano de Investigaciones Cientıficas, Caracas, Venezuela

Yrene DıazFacultad de Ciencias, Escuela de Quımica, Universidad Central de Venezuela, Caracas, Venezuela

Andreina FernandezEscuela de Bioanalisis, Facultad de Ciencias de la Salud, Universidad de Carabobo, Valencia, Venezuela

The absorption and fluorescence spectrums of four antibacte-rial quinolones, namely, ciprofloxacin, norfloxacin, enoxacin, andcinoxacin, were studied in the presence of human serum albumin(HSA). Of the three fluoroquinolones studied, ciprofloxacin andnorfloxacin were found to bind efficiently to HSA when irradiatedwith visible light, whereas the third, enoxacin, bound only moder-ately. On the other hand, cinoxacin, a nonfluorinated quinolone ofthe first generation, did not show any interaction with HSA. Thefindings were inferred by monitoring the evolution of the fluores-cence spectrums of the solutions as a function of time. A directrelationship between the capacity of the photo-induced defluori-nation to produce aryl cation intermediates, and the subsequentbinding reaction with HSA, was observed and is discussed.

Keywords Fluorescence, Human Serum Albumin, Photoallergy,Phototoxicity, Quinolones

Several antibacterial quinolones are known to induce pho-totoxic and photoallergic side effects after systemic or topicalapplication (Cardenas et al. 1991; Przybilla et al. 1990; Vargasand Rivas 1997). Photoallergy is the result of the covalent bind-ing of a photosensitizing drug, photochemical intermediate, orphotoproduct to a protein that results in the formation of a pho-toantigen (Epstein 1983; Miranda et al. 1998). In this context,clinical evidence of photosensitization disorders such as pho-totoxicity and photoallergy have been found to be associatedwith ciprofloxacin (Ferguson et al. 1988; Klecak et al. 1997);enoxacin (Izu et al. 1992; Kawabe et al. 1989); cinoxacin (Vargaset al. 1994); and norfloxacin (Martınez et al. 1998; Shelley and

Received 19 January 2003; accepted 7 March 2003.This research was supported by a grant from Fundacion Polar and

the German Embassy in Venezuela.Address correspondence to Franklin Vargas, Laboratorio de Foto-

quimica Organica, Centro de Quımica-IVIC, Apartado 21827, Caracas,1020-A, Venezuela. E-mail: fvargas@ivic.ve

Shelley 1988). Fluoroquinolones are amino acids and, as such,they exhibit amphoteric properties. At neutral pH conditions, thedominant form is zwitterionic; below pH 5 the form is cationic;and above pH 9 the form is anionic. Thus, their photochemistryis expected to be pH dependent.

Previous studies have demonstrated that the main photo-chemical process under pH-neutral conditions is defluorination(Fasani et al. 1999). It has been known that the presence of ahalogen atom at position 8 strongly enhances the phototoxicityof quinolone drugs (Marutani et al. 1993). This, coupled with theformation of highly reactive intermediates such as aryl cationsduring the defluorination process, suggests that the adverse bio-logical effect may be related to irreversible reactions involvingthe covalent bonding of this cation to some cell component.

From a biopharmaceutical point of view, one of themost important biological functions of HSA is to transportdrugs as well as other endogenous or exogenous substancesthrough cell membranes. The aim of this study was to deter-mine the binding capacity between each of four antibacterialquinolones (1) ciprofloxacin; (2) norfloxacin (a third-generationquinolones); (3) enoxacin (a second-generation quinolone); and(4) cinoxacin (a first-generation quinolone (Fig. 1)) and HSAat a neutral pH by means of fluorescence as well as absorptionspectroscopy so as to establish a simple and reliable method ofscreening photoallergens (Barratt and Brown 1985; Gonzalez-Jimenez and Garcıa-Cantalejo 2002). The quinolones studiedwere the method consisted of preparing quinolone solutions ofthe order of 10−4 M in various solvents, that is, alcoholic orphosphate-buffered saline (PBS), to keep pH values at about 7.These solutions were titrated with aliquots of an aqueous stocksolution of HSA directly into the absorbance or fluorescencecell, according to the chosen method, to determine the prop-erty under investigation. The experimental results indicated thatfluorescence determinations were a better choice than the ab-sorbance approach.

221

Tox

icol

ogy

Mec

hani

sms

and

Met

hods

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

UB

Kie

l on

11/0

9/14

For

pers

onal

use

onl

y.

222 F. VARGAS ET AL.

FIG. 1. Structures of the antibacterial quinolones studied.

MATERIALS AND METHODSAll analytical or high-performance liquid chromatogra-

phy (HPLC)–grade solvents were obtained from Merck(Darmstadt, Germany). HSA, Sephadex G-25, sodium chlo-ride (NaCl), ciprofloxacin, enoxacin, norfloxacin, and cinoxacinwere purchased from Sigma Chemical (St. Louis, MO).

The experiments were carried out using a Rayonet pho-tochemical chamber reactor (model RPR-100, Southern NewEngland Ultraviolet, Connecticut, USA) equipped with 16 phos-phorus lamps that had an ultraviolet (UV)-A emission maximumbetween 320 and 400 nm and a UV-B emission maximum be-tween 290 and 320 nm, or 23 mW/cm2 of irradiance as measuredby a UVX digital radiometer (Melles Griot, USA). The distancebetween the light source and the test aliquots was 10 cm. Thetemperatures detected in the cuvette during a standard 1-h irra-diation were no higher than 27◦C.

UV-Vis spectrophotometry of the antibacterial quinolone so-lutions was followed by using a Milton Roy Spectronic 3000array instrument (Milton Roy Company, USA). The fluores-cence spectrums were registered by a Shimadzu RF 1501 spec-trofluorophotometer. These spectrophotometers were used tocharacterize the properties of the quinolones in a PBS solution(pH 7.4, 0.01 M phosphate buffer and 0.135 M NaCl) and alsoethanol, as well as to monitor the titration with HSA before andafter the irradiation.

The titration of the quinolone solutions (1.0 × 10−4 M)with HSA was performed by adding appropriate aliquots of an

aqueous-buffered HSA stock solution at a 1.0-mM concentra-tion (pH 7.4) directly to the absorbance or fluorescence cellso that the final protein concentration was in the range of 0 to5.0×10−4 M. The solutions were allowed to incubate in the darkfor 20 min. Then the samples in the 1-cm2 Suprasil quartz cellswere irradiated under the noted conditions for various lengths oftime. The control included drug protein mixtures that were alsoplaced in darkness and irradiated for the same periods of time.The drug was separated from the protein by a Sephadex G-25column equilibrated with PBS. The photobinding was monitoredby UV-Vis and fluorescence spectroscopy.

RESULTS AND DISCUSSIONOur investigations were directed toward experimental stud-

ies of the in vitro drug-binding capacity of a model protein suchas HSA. It is possible to determine the relative ability of eachindividual drug to bind effectively to the substrate. The pho-tochemistry and the in vitro phototoxicity associated with thecompounds under investigation, often photosensitizing pharma-cological agents, provide important data for establishing themechanism of their phototoxic activity.

The courses of the photolysis reactions of the neutral solu-tions of ciprofloxacin, norfloxacin, and enoxacin titrated withHSA are shown in Figures 2, 3, and 4. A gradual decrease in anddisplacement of the fluorescence bands of the quinolones wasobserved during irradiation.

Tox

icol

ogy

Mec

hani

sms

and

Met

hods

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

UB

Kie

l on

11/0

9/14

For

pers

onal

use

onl

y.

INTERACTION OF QUINOLONES WITH HSA 223

FIG. 2. Fluorescence emission spectrums obtained after titration of HSA with aliquots of ciprofloxacin (λexc = 303 nm).

FIG. 3. Fluorescence emission spectrums obtained after titration of HSA with aliquots of norfloxacin (λexc = 356 nm).

Tox

icol

ogy

Mec

hani

sms

and

Met

hods

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

UB

Kie

l on

11/0

9/14

For

pers

onal

use

onl

y.

224 F. VARGAS ET AL.

FIG. 4. Fluorescence emission spectrums obtained after titration of HSA with aliquots of enoxacin (λexc = 384 nm).

An emission intensity was observed in the fluorescenceof nonirradiated quinolone-HSA mixtures prior to and afterSephadex filtration. The fluorescence observed in the mixturesbefore filtration was caused by the photoproducts of the irradi-ated quinolones; it almost disappears when the sample is filteredthrough the Sephadex because of the complete removal of theseproducts from the mixture (data not shown). This fact rules outthe possibility that the four quinolones and their photoproductsare tightly bound to the HSA prior to irradiation. After irra-diation, the fluorescence emission remained the same beforeand after filtration. This emission, shown by the protein fractionafter filtration through the Sephadex, demonstrates the cova-lent photobinding of ciprofloxacin, norfloxacin, and enoxacin toHSA.

It is reasonable to think that if a nucleophilic reagent is ableto attack the reactive intermediates (aryl cations) formed dur-ing the photo-induced defluorination process of the compoundsciprofloxacin, norfloxacin, and enoxacin (Fasani et al. 1998),nucleophilic terminal groups of proteins such as OH, SH, or

NH2 could just as well be able to attack these quinolones inthe same way, resulting in a covalent photobinding to the pro-tein. This process would then be responsible for the formationof a photoantigen, which is the first step in a photoallergic reac-tion, as was demonstrated for the similar quinolone fleroxacin inLangerhans-cell–enriched epidermal cells prepared from miceand exposed to UV-A light (Ohshima et al. 2000). These cellsserve as photoantigen-presenting cells in drug photoallergies.However, if the drug has a fluorescent chromophore, that couldbe used as a less elaborate and equally efficient test for photo-

binding, and it has the additional advantage of having a highersensitivity (Barratt and Brown 1985; Moser et al. 2000). A com-mon test for photobinding involves monitoring the changes influorescence emission upon irradiation of a drug-protein mix-ture. In this case, fluorescence measurements provided clear ev-idence of drug-protein photobinding. The emission maximumsof the fluorescence spectrums of the mixtures before and afterirradiation showed important differences. The fluorescence ofciprofloxacin and norfloxacin (λmax = 339 nm, and 340 nm, re-spectively) and the photosensitized reaction product with HSAshowed a displacement toward longer wavelengths, rangingfrom 415 to 453 nm, where it finally begins a continual decreasein intensity during irradiation (see Figs. 2 and 3). Enoxacin,with a fluorescence emission of 337 and 598 nm, showed a de-crease in these bands but no displacement to longer wavelengths(see Fig. 4).

It is interesting that a quinolone such as cinoxacin that doesn’tpossess a fluorine atom in its structure shows no remarkablephotoreaction with HSA (Fig. 5); no doubt this is because nocationic structure is formed after defluorination.

The titrations were followed by means of UV-Vis spectropho-tometry as well. This technique was less sensitive than fluo-rescence for measuring the capacity of the quinolones studiedto photobind HSA. Ciprofloxacin, norfloxacin, and enoxacinshowed displacements in their absorption-band maximums afterbeing irradiated in presence of HSA and treated by the Sephadex:from 276 to 280 nm in ciprofloxacin; from 278 to 281 nm innorfloxacin; and from 340 to 345 nm in enoxacin. However,cinoxacin didn’t show any detectable change. During titration

Tox

icol

ogy

Mec

hani

sms

and

Met

hods

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

UB

Kie

l on

11/0

9/14

For

pers

onal

use

onl

y.

INTERACTION OF QUINOLONES WITH HSA 225

FIG. 5. Fluorescence emission spectrums obtained after titration of HSA with aliquots of cinoxacin (λexc = 305 nm).

in darkness, no change was observed in the absorbance maxi-mums of the four compounds.

CONCLUSIONSThe covalent photobinding of the fluorinated quinolone an-

tibiotics ciprofloxacin, norfloxacin, and enoxacin to a model

FIG. 6. Mechanisms of carbocation and carbene mesomer formation after photolysis of ciprofloxacin, norfloxacin, and enoxacin.

protein such as HSA was studied by means of fluorescence spec-troscopy. Photo-induced defluorination of the fluoroquinolonesunder investigation generated a reactive cationic intermediate(Fig. 6) that readily bound to nucleophilic groups in the proteinmolecule. This was demonstrated by monitoring the evolutionof the fluorescence spectrums, namely, a decrease in emission

Tox

icol

ogy

Mec

hani

sms

and

Met

hods

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

UB

Kie

l on

11/0

9/14

For

pers

onal

use

onl

y.

226 F. VARGAS ET AL.

intensity and the displacement of the fluorescence band to longerwavelengths, from 276 to 280 nm for ciprofloxacin, from 278 to281 nm for norfloxacin and from 340 to 345 nm for enoxacin.No changes in the spectrums were detected during titration indarkness. Since these reactions are presumably the first steps thatoccur at the inception of a photoallergy, this investigation maycontribute to a better understanding of adverse skin reactionsproduced by other families of drugs on exposure to sunlight orartificial sources of UV-A light.

REFERENCESBarratt, M. D., and Brown, K. R. 1985. Photochemical binding of photoallergens

to human serum albumin: a simple in vitro method for screening potentialphotoallergens. Toxic. Lett. 24:1–6.

Cardenas, A. M., Vargas, F., Fernandez, E., and Hidalgo, M. E. 1991. Pho-totoxic potential of quinolones. J. Photochem. Photobiol. B: Biol. 10:249–255.

Epstein, J. H. 1983. Phototoxicity and photoallergy in man. J. Am. Acad. Der-matol. 8:141–147.

Fasani, E., Albini, A., Mella, M., Rampi, M., and Negra, F. B. 1999. Light anddrugs: the photochemistry of fluoroquinolone antibiotics. Int. J. Photoenergy1:7–11.

Fasani, E., Profumo, A., and Albini, A. 1998. Structure and medium-dependentphotodecomposition of fluoroquinolone antibiotics. Photochem. Photobiol.68:666–674.

Ferguson, J., McIntosh, J., and Walker, E. M. 1988. Ciprofloxacin-inducedphotosensitivity: in vitro and in vivo studies. J. Invest. Dermatol. 91:385.

Gonzalez-Jimenez, J., and Garcıa-Cantalejo, J. 2002. Interactions of drugs withhuman serum albumin studied by fluorescente spectroscopy. In Advances inBiomedical Applications of Photochemistry and Photobiology, ed. F. Vargas,51–78. Trivandrum, India: Research Signpost.

Izu, R., Gardeazabal, J., Gonzalez, M., Landa, N., Raton, J. A., and Dıaz-Perez,J. L. 1992. Enoxacin-induced photosensitivity: study of two cases. Photoder-matol. Photoimmunol. Photomed. 9:86–88.

Kawabe, Y., Mizuno, N., and Sakakibara, S. 1989. Photoallergic reactions causedby enoxacin. Photodermatol. Photoimmunol. Photomed. 6:58–60.

Klecak, G., Urbach, F., and Urwyler, H. 1997. Fluoroquinolone antibacterialsenhance UVA-induced skin tumors. J. Photochem. Photobiol. B: Biol. 37:174–181.

Martınez, L., Sik, R., and Chignell, C. 1998. Fluoroquinolone antimicrobials:singlet oxygen, superoxide and phototoxicity. Photochem. Photobiol. 67:399–403.

Marutani, K., Matsumoto, M., Otabe, Y., Nagamuta, M., Tanaka, K., Miyoshi,A., Hasegawa, T., Nagano, H., Matsubara, S., Kamide, R., Yokota, T.,Matsumoto, F., and Ueda, Y. 1993. Reduced phototoxicity of a fluoro-quinolone antibacterial agent with a methoxy group at the 8 position in miceirradiated with long-wavelength UV light. Antomicrob. Agents Chemother.37:2217–2223.

Miranda, M. A., Castell, J. V., Hernandez, D., Gomez-Lechon, M. J., Morera,I. M., and Sarabia, Z. 1998. Drug photosensitized protein modification: iden-tification of the reactive sites and elucidation of the reaction mechanism withtiaprofenic acid/albumin as a model system. Chem. Res. Toxicol. 11:172–177.

Moser, J., Bosca, F., Lovell, W. W., Castell, J. V., Miranda, M. A., and Hye,A. 2000. Photobinding of carprofen to protein. J. Photochem. Photobiol. B:Biol. 58:13–19.

Ohshima, A., Seo, N., Takigawa, M., and Tokura, Y. 2002. Formation of antigenicquinolone photoadducts on Langerhans cells initiates photoallergy to system-atically administered quinolone in mice. J. Invest. Dermatol. 114:569–575.

Przybilla, B., Georgii, A., Bergner, T., and Ring, J. 1990. Demonstration ofquinolone phototoxicity in vitro. Dermatologica 181:98–103.

Shelley, E. D., and Shelley, W. B. 1988. The subcorneal pustular drug eruption:an example induced by norfloxacin. Cutis 42:24–27.

Vargas, F., Rivas, C., and Canudas, N. 1994. Photosensitized lipid peroxidationby cinoxacin and its photoproducts: involvement of a derived peroxide in itsphototoxicity. Die Pharmazie 49:742–745.

Vargas, F., and Rivas, C. 1997. Mechanistic studies on phototoxicity induced byantibacterial quinolones. Toxic Subst. Mech. 16:81–85.

Tox

icol

ogy

Mec

hani

sms

and

Met

hods

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

UB

Kie

l on

11/0

9/14

For

pers

onal

use

onl

y.