USE OF METABOLIC INHIBITORS TO INVESTIGATE THE EXCISION REPAIR OF PYRIMIDINE DIMERS AND NON-DIMER DNA DAMAGES INDUCED IN HUMAN AND ICR 2A FROG CELLS BY SOLAR ULTRAVIOLET RADIATION

Abstract— ICR 2A frog and normal human skin fibroblasts were exposed to either 5 J/m² of 254 nm UV or 50 kJ/m² of the Mylar-filtered solar UV wavelengths produced by a fluorescent sunlamp. Following these approximately equitoxic treatments, cells were incubated in medium containing the DNA synthesis inhibitors hydroxyurea (HU) and 1–β-D-arabinofuranosyl cytosine (ara C) for 0–20 min (human fibroblasts) or 0–4 h (frog cells) to accumulate DNA breaks resulting from enzymatic incision during excision repair. It was found that breaks were formed in human cells at about a 200-f-old higher rate compared with the ICR 2A cells indicating a relatively low capacity for excision repair in the frog cells. In addition, the rate of DNA break formation in solar UV-irradiated cells was only one-third of the level detected in 254 nm-irradiated cells. This result is consistent with the conclusion that the pathway(s) involved in the repair of solar UV-induced DNA damages differs from the repair of lesions produced in cells exposed to 254 nm UV.     

Induction of DNA strand breaks and DNA-protein cross-links in normal human skin fibroblasts following exposure to 254 nm UV radiation

Levels of DNA strand breaks and DNA-protein cross-links (DPCs) were measured using the alkaline elution assay in normal human skin fibroblasts irradiated with 0-200 J m-2 of 254 nm UV radiation and incubated for 0-24 h. On incubation, the yields of both single-strand breaks (SSBs) and DPCs increased with similar kinetics and remained elevated. In addition, when SSBs were measured under conditions in which DPCs were not eliminated by treatment with proteinase K, a measurable yield of SSBs could not be detected. Hence, the SSBs that form in the UV-irradiated cells following incubation appear to be associated with the DPCs.     

Photoreactivating Enzyme for (6–4) Photoproducts in Cultured Goldfish Cells

We previously reported that when cultured goldfish cells are illuminated with fluorescent light, photorepair ability for both cyclobutane pyrimidine dimers and (6–4) photoproducts increased. In the present study, it was found that the duration of the induced photorepair ability for cyclobutane pyrimidine dimers was longer than that for (6–4) photoproducts, suggesting the presence of different photolyases for repair of these two major forms of DNA damage. A gel shift assay was then performed to show the presence of protein(s) binding to (6–4) photoproducts and its dissociation from (6–4) photoproducts under fluorescent light illumination. In addition, at 8 h after fluorescent light illumination of the cell, the binding of pro-tein(s) to (6–4) photoproducts increased. The restriction enzymes that have recognition sites containing TT or TC sequences failed to digest the UV-irradiated DNA pho-toreactivated by using Escherichia coli photolyase for cyclobutane pyrimidine dimers, indicating that restriction enzymes could not function because (6–4) photoproducts remained in recognition sites. But, when UV-irradiated DNA depleted of cyclobutane pyrimidine dimers was incubated with extract of cultured goldfish cells under fluorescent light illumination, it was digested with those restriction enzymes. These results suggested the presence of (6–4) photolyase in cultured goldfish cells as in Dro-sophila, Xenopus and Crotalus.     

ULTRAVIOLET INACTIVATION OF THE ABILITY OF E. COLI TO SUPPORT THE GROWTH OF PHAGE T7: AN ACTION SPECTRUM

Abstract— Ultraviolet (UV)-irradiated E. coli K-12 wild-type cells were sensitized by a post-irradiation treatment with 10^-2 M 2, 4-dinitrophenol (DNP). This effect was not seen in strains carrying a uvr mutation, suggesting that DN P interferes with the excision repair process. The polA strain was sensitized to the same extent as the wild-type strain, while the exrA strain was not affected by DNP treatment.
Recombination deficient strains ( recA, recB and recA recB ) were protected by DNP treatment after UV irradiation. This protection was abolished by the addition of a uvr mutation (i.e., in strains recA uvrB and recB uvrB ).
Alkaline sucrose gradient sedimentation studies showed that DNP treatment interfered with the rejoining of DNA single-strand breaks induced by the excision repair process. This interference was apparently specific for the exr gene-dependent branch of the uvr gene-dependent excision repair process, since the uvr and exr strains were not sensitized while the wild-type and polA strains were sensitized.     

PRIMARY DNA DAMAGE, HPRT MUTATION AND CELL INACTIVATION PHOTOINDUCED WITH VARIOUS SENSITIZERS IN V79 CELLS

DNA strand breaks and hypoxanthine guanine phosphoribosyl transferase (HPRT) mutants were measured in parallel in photochemically treated (PCT) cells and compared at the same level of cell survival. Chinese hamster fibroblasts (V79 cells) were either incubated with the lipophilic dyes tetra(3—hydroxyphenyl)porphyrin (3THPP) and Photofrin II (PII), the anionic dye meso -tetra(4—sulfonatophenyl)porphine (TPPS₄) or the cationic dye meso -tetra( N -methyl-4-pyridyl)porphine ( p -TMPyPH₂ before light exposure. In the cells, the lipophilic dyes were localized in membranes, including the nuclear membrane, while the hydrophilic dyes were taken up primarily into spots in the cytoplasm. In addition, the hydrophilic TPPS₄ was distributed homogeneously throughout the whole cytoplasm and nucleoplasm. According to the HPRT mutation test, the mutagenicity of light doses survived by 10% of the cells was a factor of six higher in the presence of 3THPP than of PII, whereas for X-rays it was a factor of three higher than for PCT with 3THPP. Light exposure in the presence of the hydrophilic dyes TPPS₄ and p -TMPyPH₂ was not significantly mutagenic. There was no correlation between the induced rates of HPRT mutants and of DNA strand breaks. Thus, TPPS₄ was the most efficient sensitizer with regard to DNA strand breaks when compared at the same level of cell survival, followed by 3THPP, PII and p -TMPyPH₂. Hence, the rate of DNA strand breaks cannot be used to predict the mutagenicity of PCT.     

EFFECTS OF CAFFEINE ON DNA REPAIR OF UV-IRRADIATED DICTYOSTELIUM DISCOIDEUM

Abstract— Caffeine enhances the UV-killing of amoeboid cells of NC-4, but UV-irradiated γs-13 is insensitive to caffeine. UV-irradiated NC-4 becomes insensitive to the effect of caffeine during a postirradiation incubation in buffer for about 90 min, but γs-13 remains unchanged in the sensitivity to caffeine throughout the incubation for 180 min. Amoeboid cells of γs-13 can remove pyrimidine dimers as well as NC-4 even in the presence of caffeine. Caffeine inhibits rejoining of strand-breaks of DNA in UV-irradiated NC-4, but the rejoining in γs-13 is insensitive to caffeine.     

Effect of glutathione on lambda deoxyribonucleic acid strand breaks in the reaction system of glutathione-alloxan in the presence of Fe(3+)-ethylenediaminetetraacetic acid.

Alkaline sucrose density gradient and agarose gel electrophoresis methods were used to observe lambda deoxyribonucleic acid (DNA) strand breaks by the reaction system of reduced glutathione (GSH) with alloxan in the presence of Fe(3+)-ethylenediaminetetraacetic acid (EDTA). When DNA was incubated in the reaction system for 10 min, DNA strand breaks were easily induced. The increasing concentrations of GSH up to 1.0 mM in the reaction system in the presence of 1.0 mM alloxan caused DNA strand breaks in a concentration-dependent fashion and GSH beyond 2.0 mM caused in the strand breaks of DNA by which the fragments with multiple ranges of molecular weight were produced. The strand breaks of DNA in the reaction system containing low concentrations of GSH were protected by catalase and hydroxyl radical (HO.) scavengers but superoxide dismutase (SOD) did not, indicating that such breaks were induced by HO.generated from the Fenton reaction. On the other hand, the strand breaks of DNA at high concentrations of GSH were protected by ethanol and desferrioxamine, but not effectively by SOD and HO.scavengers, suggesting the possible participation of some oxidizing species of iron rather than HO.. These results indicate that HO.or oxidizing species of iron generated in the GSH-alloxan system depending on the concentration of GSH attacks DNA to produce strand breaks.     

DETECTION OF CYCLOBUTANE THYMINE DIMERS IN DNA OF HUMAN CELLS WITH MONOCLONAL ANTIBODIES RAISED AGAINST A THYMINE DIMER- CONTAINING TETRANUCLEOTIDE

Abstract— A hybrid cell line (hybridoma) has been isolated after fusion between mouse-plasmacytoma cells and spleen cells from mice immunized with a thymine dimer-containing tetranucleotide coupled to a carrier protein. Monoclonal antibodies produced by this hybridoma were characterized by testing the effect of various inhibitors in a competitive enzyme-linked immunosorbent assay (ELISA). The antibodies have a high specificity for thymine dimers in single-stranded DNA or poly(dT), but do not bind UV-irradiated d(TpC)₅. Less binding is observed with short thymine dimer-containing sequences. In vitro treatment of UV-irradiated DNA with photoreactivating enzyme in the presence of light, or with Micrococcus luteus UV-endonuclease results in disappearance of antigenicity. Antibody-binding to DNA isolated from UV-irradiated human fibroblasts (at 254 nm) is linear with dose. Removal of thymine dimers in these cells during a post-irradiation incubation, as detected with the antibodies, is fast initially but the rate rapidly decreases (about 50% residual dimers at 20 h after 10 J/m²). The induction of thymine dimers in human skin irradiated with low doses of UV-B, too, was demonstrated immunochemically, by ELISA as well as by quantitative immunofluorescence microscopy.     

Photogenotoxicity of Skin Phototumorigenic Fluoroquinolone Antibiotics Detected Using the Comet Assay

Abstract— The fluoroquinolone(FQ) antibiotics photosensitize human skin to solar UV radiation and are reported to photosensitize tumor formation in mouse skin. As tumor initiation will not occur without genotoxic insult, we examined the potential of ciprofloxacin, lomefloxacin, fle-roxacin, BAYy3118 (a recently developed monofluori-nated quinolone) and nalidixic acid to photosensitize DNA damage in V79 hamster fibroblasts in vitro. Cells were exposed to 37.5 kj/m₂ UVA (320-400 nm; glass filtered Sylvania psoralen + UVA (PUVA) tubes; calibrated Waldmann radiometer) at 4A_oC in the presence of FQ and immediately afterwards embedded in agarose, lysed and placed in an electrophoretic field at pH 12. Under these denaturing conditions, the presence of DNA single-strand breaks (SSB), alkali-labile sites (ALS) and double-strand breaks (DSB) can be visualized as DNA migrating away from the nucleus (characteristic "comet" appearance) after staining with a specific fluorochrome. At FQ concentrations that induced minimal loss of cell viability (neutral red uptake assay) the compounds tested induced comets with a rank order of BAYy3118 norfloxacin ciprofloxacin lomefloxacin fleroxacin nalidixic acid. If cells were incubated after treatment for 1 h at 37_oC, the comet score decreased, suggesting efficient removal of SSB/ALS/DSB. Addition of the DNA polymerase, inhibitor, aphidicolin, to cells treated with either ciprofloxacin alone or ciprofloxacin + UVA resulted in an accumulation of SSB due to the endo/exonuclease steps of excision repair. We have demonstrated that the FQ are photogenotoxic in mammalian cells but that FQ-pho-tosensitized SSB are efficiently repaired. Preliminary evidence that ciprofloxacin photosensitizes the formation of DNA lesions warranting excision repair may indicate production of more mutagenic lesions.     

Lack of correlation between DNA strand breakage and p53 protein levels in human fibroblast strains exposed to ultraviolet lights

The contribution of DNA strand breaks accumulating in the course of nucleotide excision repair to upregulation of the p53 tumor suppressor protein was investigated in human dermal fibroblast strains after treatment with 254 nm ultraviolet (UV) light. For this purpose, fibroblast cultures were exposed to UV and incubated for 3 h in the presence or absence of l-beta-D-arabinofuranosylcytosine (araC) and/or hydroxyurea (HU), and then assayed for DNA strand breakage and p53 protein levels. As expected from previous studies, incubation of normal and ataxia telangiectasia (AT) fibroblasts with araC and HU after UV irradiation resulted in an accumulation of DNA strand breaks. Such araC/HU-accumulated strand breaks (reflecting nonligated repair-incision events) following UV irradiation were not detected in xeroderma pigmentosum (XP) fibroblast strains belonging to complementation groups A and G. Western blot analysis revealed that normal fibroblasts exhibited little upregulation of p53 (approximately 1.2-fold) when incubated without araC after 5 J/m2 irradiation, but showed significant (three-fold) upregulation of p53 when incubated with araC after irradiation. AraC is known to inhibit nucleotide excision repair at both the damage removal and repair resynthesis steps. Therefore, the potentiation of UV-induced upregulation of p53 evoked by araC in normal cells may be a consequence of either persistent bulky DNA lesions or persistent incision-associated DNA strand breaks. To distinguish between these two possibilities, we determined p53 induction in AT fibroblasts (which do not upregulate p53 in response to DNA strand breakage) and in XP fibroblasts (which do not exhibit incision-associated breaks after UV irradiation). The p53 response after treatment with 5 J/m2 UV and incubation with araC was similar in AT, XPA, XPG and normal fibroblasts. In addition, exposure of XPA and XPG fibroblasts to UV (5, 10 or 20 J/m2) followed by incubation without araC resulted in a strong upregulation of p53. We further demonstrated that HU, an inhibitor of replicative DNA synthesis (but not of nucleotide excision repair), had no significant impact on p53 protein levels in UV irradiated and unirradiated human fibroblasts. We conclude that upregulation of p53 at early times after exposure of diploid human fibroblasts to UV light is triggered by persistent bulky DNA lesions, and that incision-associated DNA strand breaks accumulating in the course of nucleotide excision repair and breaks arising as a result of inhibition of DNA replication contribute little (if anything) to upregulation of p53.     

A FILTER BINDING ASSAY FOR CHARACTERIZATION OF ANTIBODIES TOWARDS DAMAGED DNA

A nitrocellulose filter binding assay was used to characterize rabbit antibodies specific for UV-irradiated DNA. The dissociation constant for the UV-DNA-antibody complex was found to be 4.2 × 10⁻¹² M. No significant binding to unirradiated DNA was observed. Unlabelled single-stranded and double-stranded DNA competed with equal efficiency for the radioactively labelled antigen (UV-irradiated φX174 DNA). The antibodies also bound to OsO₄-treated DNA, suggesting that these polyclonal antibodies also recognize minor photoproducts. Caffeine efficiently decreased binding of the antibodies to UV-irradiated DNA.     

LETHALITY AND THE INDUCTION AND REPAIR OF DNA DAMAGE IN FAR, MID OR NEAR UV-IRRADIATED HUMAN FIBROBLASTS: COMPARISON OF EFFECTS IN NORMAL, XERODERMA PIGMENTOSUM AND BLOOM'S SYNDROME CELLS

Abstract Using normal human fibroblasts we have determined the ability of far (254 nm), mid (310 nm) or near (365 nm) UV radiation to: (i) induce pyrimidine dimers (detected as UV endonuclease sensitive sites) and DNA single-strand breaks (detected in alkali); (ii) elicit excision repair, monitored as unscheduled DNA synthesis (UDS); and (iii) reduce colony-forming ability. Unscheduled DNA synthesis studies were also performed on dimer excision-defective xeroderma pigmentosum (XP) cells, and the survival studies were extended to include XP and Bloom's syndrome (BS) strains. UV-induced cell killing in normal, BS and XP cells was found to relate to an equivalent dimer load per genome after 254 or 310 nm exposure, whereas at 365 nm the lethal effects of non-dimer damage appeared to predominate. Lethality could not be correlated with DNA strand breakage at any wavelength. The two XP strains examined showed the same relative UDS repair deficiency at the two shorter wavelengths in keeping with a predominant role for pyrimidine dimer repair in the expression of UDS. However, UDS was not detected in 365 nm UV-irradiated normal and XP cells despite dimer induction; this effect was due to the inhibition of DNA repair functions since 365 nm UV-irradiated normal cells showed reduced capacity to perform UDS subsequent to challenge with 254 nm UV radiation.
In short, the near UV component of sunlight apparently induces biologically important non-dimer damage in human cells and inhibits DNA repair processes, two actions which should be considered when assessing the deleterious actions of solar UV.     

UV-DNA Damage in Mouse and Human Cells Induces the Expression of Tumor Necrosis Factor α

Ultraviolet light induces the expression of tumor necrosis factor α (TNFα) in many mammalian cells. We have examined the signal for this induction in a human DNA repair-deficient cell line carrying a transgene composed of the murine TNF regulatory sequences fused to the chloramphenicol acetyltransferase (CAT) structural gene. When compared by fluence, UVC was a more efficient inducer of CAT than was UVB, but they were equivalent inducers when compared by the frequency of cyclobutyl pyrimidine dimers produced by each source. Further, treatment of UV-irradiated cells with the prokaryotic DNA repair enzyme T4 endonuclease V in-creased the level of repair of dimers and concomitantly reduced CAT gene expression. Membrane-bound TNFα expression was increased by UV and reduced by repair of dimers. Finally, in the TNFcat transgene system, DNA damage directly to the cell with the transgene was required as cocultivation of unirradiated TNFcat cells with UV-irradiated cells did not increase CAT activity. These results show that DNA damage is a signal for the induction of TNFa gene expression in mouse and human cells.     

The soluble eumelanin precursor 5,6-dihydroxyindole-2-carboxylic acid enhances oxidative damage in human keratinocyte DNA after UVA irradiation.

Soluble melanin precursors are present in serum and may act as skin chromophores contributing to UVR-induced oxidative damage. Our study aimed to determine whether the soluble eumelanin precursor 5,6-dihydroxy-indole-2-carboxylic acid (DHICA) photosensitizes DNA damage in human keratinocytes exposed to UVA irradiation. The HaCaT keratinocytes were incubated with and without DHICA, before irradiation with broadband UVA (320-400 nm). The DNA photodamage was assessed using the comet assay that detects frank single-strand breaks (SSB) and specific oxidative lesions with the addition of endonuclease III. Without DHICA incubation, there was no significant increase in SSB, compared to unirradiated cells, for doses up to 48.5 J/cm2 (< 1 minimum erythemal dose). Preincubation with 0.5 microM DHICA caused an increase in SSB at every UVA dose (significant from 12.1 to 48.5 J/cm2), while varying the DHICA concentration (0.125-2 microM) showed this effect to be concentration dependent such that SSB increased and endonuclease III-sensitive sites decreased with increasing DHICA concentration. The irradiation of cells in the presence of antioxidants (catalase, mannitol and histidine) suggests that DHICA-induced photosensitization is mediated via singlet oxygen and, to a lesser extent, hydroxyl radicals. These results indicate that DHICA can induce strand breaks with UVA at clinically relevant doses via a mechanism involving reactive oxygen species.     

INDUCIBLE POSTREPLICATION REPAIR IS RESPONSIBLE FOR MINIMAL MEDIUM RECOVERY IN UV-IRRADIATED Escherichia coliK–12

Abstract— Ultraviolet (UV)-irradiated Escherichia coli K–12 uvrA cells showed higher survival if plated on minimal growth medium rather than on rich growth medium, i.e., they showed minimal medium recovery (MMR). A 2-hour treatment of UV-irradiated cells with rifampicin inhibited the subsequent expression of MMR, and produced a large reduction in survival. We have recently isolated a new mutant ( mmrA1 ) that does not show MMR. The mmrA mutation protected UV-irradiated uvrA cells from the effect of rich growth medium on survival, but not from the effect of rifampicin on survival. DNA daughter-strand gap (DSG) repair in UV-irradiated (4 J/m²) uvrA cells was inhibited to the same degree whether rich growth medium was added immediately after irradiation or after 10 min of postirradiation incubation in minimal growth medium. However, chloramphenicol added immediately after irradiation greatly reduced this repair; there was less reduction if it was added 10 min after UV irradiation. These findings suggest that MMR is an inducible repair phenomenon, and that rich growth medium inhibits this repair process itself rather than its induction.     

UV-ENHANCED REACTIVATION OF MINUTE-VIRUS-OF-MICE: STIMULATION OF A LATE STEP IN THE VIRAL LIFE CYCLE

Abstract— UV-enhanced reactivation of minute-virus-of-mice (MVM), an autonomous parvovirus, was studied in parasynchronous mouse A9 cells. The survival of UV-irradiated MVM is increased in cells which have been UV-irradiated prior to infection. UV-enhanced reactivation can be explained neither by facilitated plaque detection on UV-treated indicator cells, nor by altered kinetics of virus production by UV-irradiated cells. No effect of the multiplicity of infection on virus survival was detected in unirra-diated or irradiated cells. The magnitude of UV-enhanced reactivation is a direct exponential function of the UV dose administered to the virus while virus survival is inversely proportional to the UV dosage. The expression of UV-enhanced reactivation can be activated in cells arrested in G₀, it requires de novo protein synthesis and it is maximal when cells are irradiated 30 h before the onset of viral DNA replication. Early phases of the viral cycle, such as adsorption to cellular receptors, migration to the nucleus and uncoating, were not affected by cell irradiation and are unlikely targets of the UV-enhanced reactivation function(s). These results, together with the single-strandedness of the viral genome, strongly suggest that the step stimulated in UV-irradiated cells functions concomitant with, or subsequent to, viral DNA replication.     

Effects of photo-activated bleomycin on deoxyribonuclease I, exonuclease III and deoxyribonucleic acid polymerase I reactions

The activity of bleomycin to break the strand of deoxyribonucleic acid (DNA) in the presence of 2-hydroxy-1-ethanethiol (2-mercaptoethanol) was enhanced by ultraviolet (UV) irradiation. Photo-activated bleomycin stimulated the action of deoxyribonuclease I (DNase I) to degrade DNA and the DNA synthesis by DNA polymerase I with DNase I. On the other hand, although UV-irradiated bleomycin scarcely broke the DNA strand in the presence of 1,2-benzenediol (catechol), it stimulated the action of DNase I to degrade DNA in the presence of catechol. In accordance with the inhibition by catechol, when DNA treated with UV-irradiated bleomycin in the presence of catechol was employed as a primer for the DNA synthesis, the incorporation of precursor into the acid-insoluble fraction by DNA polymerase I with exonuclease III was reduced to about one-half of the incorporation into DNA treated with unirradiated bleomycin. These findings suggest that the ability of bleomycin to bind to double-helical DNA forming regions sensitive to DNase I was increased by an appropriate dose of UV irradiation and that catechol inhibited the activity of the UV-irradiated bleomycin to break the DNA strand rather than to bind to DNA.     

CHARACTERISATION OF TWO MONOCLONAL ANTIBODIES SPECIFIC FOR DIMERISED AND NON-DIMERISED ADJACENT THYMIDINES IN SINGLE STRANDED DNA*

Abstract— Hybrid cell lines (hybridomas) have been isolated from fusions between P3-NS1-1-Ag4-1 mouse myeloma cells and spleen cells from BALB/c mice hyperimmunised with UV-irradiated single-stranded DNA (UVssDNA) and UV-irradiated polydeoxythymidylic acid (UVpolydT). Monoclonal antibodies from two different hybridomas are characterised in the present report by competitive inhibition with different synthetic polynucleotides and oligonucleotides. The first antibody, designated αUVssDNA-1, recognises thymidine dimers in a polynucleotide or an oligonucleotide sequence at least four nucleotides long but not isolated thymidine dimers, suggesting that it recognises the conformational change associated with thymidine dimers. The second antibody, designated αssDNA-2, recognises unirradiated or UV-irradiated tracts of thymidine, but will not crossreact with tracts of other nucleotides (A, G, C, A.T, G.C, C.U, U). Inhibition of binding of αUVssDNA-1 to [³H]-UVssDNA by calf thymus UVssDNA is dependent on UV exposure and wavelength as expected from the antigenic determinant.     

BIOLOGICAL PROPERTIES OF SOME BENZOPSORALEN DERIVATIVES

Abstract— The biological activity of some benzopsoralen derivatives, prepared with the aim of obtaining new drugs for photochemotherapy, has been studied. The more interesting compounds are 4-hydroxy-methyl-4',5'-benzopsoralen and 4-hydroxymethyl-4',5'-tetrahydro-benzopsoralen, which were found to be active in the dark also: DNA and RNA synthesis were both inhibited in Ehrlich cells, even if in a partially reversible fashion, while protein synthesis remained unaffected. In Chinese hamster ovary cells cultured in vitro , the clonal growth was strongly inhibited by incubation in the dark with both drugs, while a number of chromosomal aberrations was observed in the fraction of growing cells. Using alkaline elution, DNA strand breaks were detected. In addition, in the presence of aphidicolin, a specific inhibitor of DNA polymerase, the clonal growing capacity was completely restored; in contrast, the number of DNA strand breaks remained unchanged. All these results suggest that DNA topoisomerases are probably the target of these two benzopsoralens. These compounds are also good sensitizers; by UV-A irradiation they have a good capacity to produce singlet oxygen, but they appeared to be unable to induce erythemas on guinea-pig skin. Under UV-A light, they induced a strong inhibition of DNA synthesis in Ehrlich cells. Thus, benzopsoralens appear to be capable of inducing strong antiproliferative effects by two different mechanisms, by UV-A irradiation and in the dark.