Photoinduced Damage to Cellular DNA: Direct and Photosensitized Reactions

Authors


  • This paper is part of the Special Issue in Commemoration of the 70th birthday of Dr. David R. Bickers.

Corresponding author emails: jean.cadet@cea.fr (Jean Cadet) and thierry.douki@cea.fr (Thierry Douki)

Abstract

The survey focuses on recent aspects of photochemical reactions to cellular DNA that are implicated through the predominant formation of mostly bipyrimidine photoproducts in deleterious effects of human exposure to sunlight. Recent developments in analytical methods have allowed accurate and quantitative measurements of the main DNA photoproducts in cells and human skin. Highly mutagenic CC and CT bipyrimidine photoproducts, including cyclobutane pyrimidine dimers and pyrimidine (6-4) pyrimidone photoproducts (6-4PPs) are generated in low yields with respect to TT and TC photoproducts. Another striking finding deals with the formation of Dewar valence isomers, the third class of bipyrimidine photoproducts that is accounted for by UVA-mediated isomerization of initially UVB generated 6-4PPs. Cyclobutadithymine (T<>T) has been unambiguously shown to be involved in the genotoxicity of UVA radiation. Thus, T<>T is formed in UVA-irradiated cellular DNA according to a direct excitation mechanism with a higher efficiency than oxidatively generated DNA damage that arises mostly through the Type II photosensitization mechanism. C<>C and C<>T are repaired at rates intermediate between those of T<>T and 6-4TT. Evidence has been also provided for the occurrence of photosensitized reactions mediated by exogenous agents that act either in an independent way or through photodynamic effects.

Introduction

Occupational and recreational exposure of humans to UVB and UVA radiations present in sunlight represents a major environmental risk factor for the development of skin cancers. Overall cumulated dose is a good predictor of induction of squamous cell carcinoma and lip cancers (1). In contrast, basal cell carcinoma and malignant melanoma incidence appears to be mostly related to intermittent exposure to solar light during childhood and adolescence (1), although the situation for melanoma appears to be more complex (2). In particular, the action spectrum is still a matter of debate. It has been recently proposed that UVB radiation is an initiator of melanoma whereas both UVB (290–315 nm) and UVA components (315–400 nm), a Class I carcinogen (3), are implicated in the progression of the disease (4). The increasing use of artificial UV sources in tanning salons also constitutes a major health issue in terms of skin cancer occurrence (5,6). It is now well documented that DNA is the main critical cellular macromolecule for determining the carcinogenic effects of solar radiation (7–9). This received strong support from the delineation of mutagenic effects of UVB and UVA components (10–14) with the observation of UVB-induced C→T transition and CC→TT mutations at bipyrimidine sites in p53 genes of skin tumors (8,15). It has been also shown that C→T transitions are generated in human skin by UV radiation within the 310–340 nm range at 5-methylcytosine containing bipyrimidine sites in CpG rich sequences (10,16). Another indirect evidence for the implication of photoinduced damage to DNA in carcinogenicity is provided by the high incidence of skin tumors in most of the complementation groups A–G of xeroderma pigmentosum patients (17,18) who suffer from deficiencies in nucleotide excision repair (19). Understanding the biological role of UV-induced DNA damage requires an accurate determination of the nature and frequency of the photoproducts. Early insights into DNA photoreactions have been gained more than 50 years ago with the isolation and characterization of the cissyn isomer of the cyclobutadithymine (T<>T) (20) in UVC-irradiated DNA (21). T<>T has been shown later to be the predominant DNA photoproduct induced in cellular DNA upon UVB and solar irradiation (22). This major discovery has provided a strong impetus to the development of research activities in the field of genotoxicity and DNA repair (23,24). Another early relevant finding dealt with the isolation of a second thymine dimeric photoproduct in UVC-irradiated thymidylyl-(3′-5′)-thymidine (TpT) (25) that has been identified as a pyrimidine (6-4) pyrimidone photoproduct (6-4PP) (26,27). This was complemented more recently by the identification of a third class of bipyrimidine photoproducts that was also isolated upon UVC irradiation of TpT and assigned as the related Dewar valence isomer (DEW) of 6-4PP (28). Other earlier striking findings concerning UVA-induced DNA photoproducts dealt with the identification of cyclobutane pyrimidine dimers (CPDs) (29,30) and oxidatively generated damage (30–32), including 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodGuo) (31) in cellular DNA. Evidence was also provided for the occurrence of photosensitization reactions mediated by exogenous compounds that are able, upon UVA excitation, to give rise to different types of DNA photoproducts, including psoralen mono- and bi-adducts to DNA. Earlier aspects of photochemical reactions involving DNA and model compounds were reported and critically discussed in several comprehensive book chapters (7,33) and review articles (22,34–40). In the present survey, emphasis is placed on achievements, mostly made during the last decade, concerning the direct and photosensitized formation of DNA damage in cells and human skin. This has been made possible by the development of accurate and quantitative methods that involve in most cases the association of high-performance liquid chromatography with electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS).

Formation of bipyrimidine photoproducts

More than half a century after the discovery of T<>T (20), a large amount of information on the basic photochemistry of DNA, either in model systems, isolated DNA or in cells, is available. The abundance of published material on the topic does not allow quoting all the excellent publications.

UVB-induced CPDs and 6-4PPs

Structure of the pyrimidine dimeric photoproduct.  DNA damage induced by UVB radiation is dominated by bipyrimidine photoproducts (22). Other photoproducts such as cytosine hydrates (41,42) or adenine-containing dimers (43,44) have been isolated in model systems but their yields are 2 orders of magnitude lower than that of bipyrimidine photoproducts in double-stranded (ds) DNA. The formation of oxidation products, more specifically 8-oxodGuo, is also induced in cellular DNA upon exposure to UVB radiation however in a low yield that is 2–3 orders of magnitude lower than that of either CPDs or 6-4PPs (32).

Two main categories of bipyrimidine photoproducts have been identified so far in UVB-irradiated DNA (Fig. 1). The first class consists of the well-known CPDs that arise from a [2+2] cycloaddition reaction between the C5–C6 double bonds of two pyrimidines. This reaction takes place in a few picoseconds as shown by recent time-resolved measurement in single-stranded DNA (45). An important question remains the nature of the excited states giving rise to CPDs. Although long believed to be only excited triplet states, growing amounts of data are accumulating to suggest a significant contribution of singlet states. Although interstrand photoproducts can be obtained under very specific conditions within isolated DNA (46), the biologically relevant CPDs involve adjacent bases. The stereochemistry of the resulting dimeric photoproduct is governed by structural features and constraints imposed by the structure of DNA duplex in native B conformation. The two bases involved in the photoinduced formation of the cyclobutane dimer show a parallel orientation and are located on the same side with respect to the cyclobutane ring. Therefore, the CPD that is exclusively generated exhibits a cis–syn stereochemistry. The four main possible CPDs that correspond to the TT, CT, TC and CC bipyrimidine sequences have been isolated in both naked and cellular DNA.

Figure 1.

 Chemical structures of the UVB-induced thymine dimeric photoproducts. CPD = cyclobutane pyrimidine dimer; 6-4PP = (6-4) photoproduct; DEW = Dewar valence isomer.

The second category of bipyrimidine photoproducts is represented by 6-4PPs. Like CPDs, the UV-mediated formation of 6-4PPs has been observed, at least for model systems containing TT, TC, CT and CC bipyrimidine sequences. 6-4PPs are produced by a [2+2] cycloaddition between the C5–C6 double bond of the 5′-end base and the C4 carbonyl group of a 3′-end thymine, the so-called Paternò-Büchi reaction. This reaction in DNA is mediated via singlet excited state as shown by the complete absence of formation of 6-4PPs upon photosensitized triplet state energy transfer. The same reaction is possible with a 3′-end cytosine in which the 4-amino group adopts an imine tautomeric form. The products of the photocycloaddition reactions are either oxetanes or azetidines, depending on whether the 3′-end base is a thymine or a cytosine. The latter highly unstable intermediates are converted into related 6-4PPs that exhibit a 3′-end pyrimidone ring and in which the C4 substituent of the 3′-end base has migrated to the C5 position of the pyrimidine moiety.

Biological consequences of the formation of bipyrimidine photoproducts cannot be detailed here. However, it may be mentioned that transfection of cells and mice with photolyases that specifically remove either CPDs or 6-4PPs indicated that the former types of lesions are responsible for the vast majority of mutagenic events induced by UV irradiation (47,48). This is explained both by the miscoding properties of deaminated cytosine in CPDs that leads to T→C transversions and CC→TT tandem mutations, and by the fact that 6-4PPs, although highly mutagenic, are efficiently removed by repair systems. Interestingly, the rate of repair has been shown to be modulated at some sequences. HPLC-ESI-MS/MS measurement also revealed that the four CPDs were not repaired with the same rate, being enzymatically removed in the following increasing order of efficiency: T<>T < T<>C < C<>CC<>T (49).

Main properties of the photoproducts.  Cyclobutane pyrimidine dimers and 6-4PPs are subject to secondary transformations. A first example is given by the photoinduced conversion of 6-4PPs into a third class of dimeric photoproducts (Fig. 1) known as the DEWs (50). The formation of these photoproducts results from the presence of a pyrimidone moiety in 6-4PPs that may absorb photons with a maximal efficiency around 320 nm. This allows a high quantum yield 4π electrocyclization reaction that is a slow process as inferred from recent time-resolved studies (51). Surprisingly, it was also reported that the 6-4TT lacking the internal phosphodiester bond is not able to undergo photorearrangement into the related Dewar valence isomer upon white light irradiation, suggesting that the DNA backbone could favor the photoisomerization of 6-4PPs. This contrasts to earlier observations showing that thymidine and thymine 6-4PPs were efficiently converted into the corresponding Dewar valence isomers upon exposure at 320 nm (52). There is a need to resolve these apparent discrepancies that may have some explanation in the different sources of UV lights used in the experiments. The 6-4PP rearrangement to Dewar valence isomers in ds DNA can be triggered by UVB radiation with however a low efficiency that requires exposure to biologically irrelevant high doses (53). This is explained by the preferential absorption of UVB photons by normal bases present in orders of magnitude larger in amounts than 6-4PPs in cells and skin. This shielding process is no longer efficient in the UVA range because normal bases poorly absorb these photons in contrast to 6-4PPs. As a result, exposure of ds DNA and cells to either simulated sunlight (SSL), a combination of UVA and UVB, or sequential exposure to UVB and UVA leads to the photoisomerization of a large fraction of 6-4PPs into DEWs (54–56).

A second major reaction that concerns the three types of bipyrimidine dimers is a hydrolytic process that converts 5,6-dihydrocytosine moieties into 5,6-dihydrouracil residues. The deamination reaction results from the substitution of the C4 amino group of cytosine by a hydroxyl function that generates a tautomeric form of uracil. Deamination takes place with unmodified cytosine but at a much slower rate. In contrast, the reaction is much more efficient when the C5–C6 double bond is saturated as is the case in CPDs and for the 5′-end bases of 6-4PPs and DEWs (57). The half reaction time is in the range of a few hours for isolated photoproducts and slightly less in ds DNA. Recently, it was shown that the efficiency of the reaction can be modulated in cells by processes such as stalling of RNA polymerase at CPD sites during transcription (58).

Distribution of photoproducts in DNA.  Quantitative information on the UV-induced formation of the three types of bipyrimidine photoproducts in cellular DNA has been gathered from immunological measurements and chromatographic analyses. Both approaches allowed the determination of the ratio between the yield of CPDs and 6-4PPs following UVB irradiation. The values reported range between 3 and 5 (53,55,59). A more extensive description of the bipyrimidine photoproduct distribution was obtained using radioactivity, fluorescence and mass spectrometry detection. For example, HPLC-ESI-MS/MS analysis showed that UVB irradiation of mammalian cells (53,54,56,60) and human skin (61) induces T<>T as the most frequent photoproduct. The corresponding 6-4TT is produced in a 10-fold lower yield. The TC sequences also show fair reactivity with both T<>C and 6-4TC being produced in similar yields. In contrast, CT and CC sites are much less susceptible to UVB-mediated reactions that lead mostly to the formation of CPDs. The corresponding 6-4PPs were detected only in small amounts in isolated DNA exposed to rather large doses of UVB. A similar CPD distribution was observed upon exposure to SSL that consists at the most 5% UVB in UVA (54). In contrast, the frequency of 6-4PPs decreases and DEWs can be detected in significant amounts while they are not present when cells are only exposed to UVB (Fig. 2).

Figure 2.

 Relative distribution of TT and TC dimeric photoproducts in Chinese hamster ovary cells exposed to either UVB or simulated sunlight (SSL). DEW isomers are observed only under the latter conditions. Photoproducts were quantified by HPLC-ESI-MS/MS (54).

In terms of absolute yields of lesions, calibrated immunoassays, electrophoretic analysis and chromatographic measurements lead to values ranging between 0.2 and 0.5 per 106 bases per J m−2 upon exposure of cultured cells to UVB (55,61,62). Variations in photoproduct frequencies observed in reported studies can be explained not only by differences in analytical approaches but also by the emission spectrum of the UVB sources used. The yield of bipyrimidine photoproducts is much lower in skin than in cells, although the distribution of photoproducts is similar (61). This result can be explained by the presence of strong UVB chromophores, especially in the stratum corneum. Interestingly, immunohistochemical analyses have shown that, for UV radiation sources with wavelengths higher than 300 nm, the frequency of CPDs and 6-4PPs remains constant in all the layers of the epidermis 63–65. The yield of pyrimidine dimers in skin is also strongly modulated by the phototype—dark skin accumulating lower amounts of lesions than fair ones (66,67).

The distribution of DNA damage described above is the average over the entire genome. It should be kept in mind that the photoreactivity of DNA strongly depends on the sequence showing high specific sites. This type of information has mainly been provided by sequencing techniques such as ligation-mediated PCR. It was shown for instance that the frequency of CPDs was not the same at all sites in a gene like p53, with ratios varying for more than 10 between the less and the most damaged ones (47). Interestingly, the sites of high photoreactivity were also found to be mutational hotspots. Sequence effects were also observed at binding sites of transcription factors. Such interactions were indeed shown in vitro to inhibit the formation of photoproducts at some positions and to enhance it at others (68). Structural parameters have also to be taken into account, such as the nucleosome positioning that modulates the photoreactivity and therefore the ratio formation between 6-4PPs and CPDs (69). Evidence has also been provided for the enhanced efficiency of photoreactions in telomeres (70) with the possible formation of interstrand lesions in the quadruplex region (71).

A number of studies have also emphasized the high photoreactivity of 5-methylcytosine (5mCyt), a minor base playing a major role in epigenetic regulation of gene expression. Sequencing techniques showed that methylated CpG sites were more prone to photoproduct formation (72) and were mutational hotspots, likely as the result of deamination (73). These results are in agreement with data obtained on model systems where the photoreactivity of 5mCyt was compared to that of unmodified cytosine (74–76). Another interesting observation was a reduced deamination rate of the methylated photoproducts. The situation might be more complex in cellular DNA where binding of specific proteins was found to decrease the deamination rate (77) while the presence of a 5′-end G enhances the reaction (78). Deamination of 5mCyt containing CPDs is also greatly modulated by rotational positioning in nucleosomes (79).

UVA-induced CPDs

UVB radiation is not the only component of the solar spectrum that leads to the formation of bipyrimidine photoproducts. In fact, UVA radiation has been shown since the 1970s, first in bacteria (29) and then in mammalian cells (32,55,80–84) and human skin (61,65,85–87), to be also able to induce CPDs. However, these observations have been poorly considered until recently. Unfounded suspicion of UVB contamination of the lamps used may explain this fact. In the last decade, a number of analytical approaches were used to quantitatively compare the effect of UVA and UVB radiations in terms of bipyrimidine photoproduct formation. Determination of action spectra and use of well-controlled UVA sources involving UVB filters provided further support to earlier observations. It is now well documented that UVA irradiation gives rise to CPDs in both cultured cells and human skin. Although tiny amounts of 6-4PPs have been detected by immunological assays (55), neither the formation of this class of damage nor of DEWs has been observed using more specific chromatographic techniques, even at high doses of UVA (46,61,88). As a striking information, the yield of CPDs was found to be larger than that of oxidatively generated lesions such as 8-oxoGua (32,46,61,88) that were long thought to be the hallmark of UVA genotoxicity, as further discussed below. The ratio between the yields of CPDs and 8-oxoGua is around 5, slightly depending on the cell type. There is however an exception for melanocytes where a CPD/8-oxoGua ratio 1.4 reflects the occurrence of a larger oxidative stress (89). Other evidence for a predominant role of CPDs in UVA genotoxicity came from mutagenesis studies in cultured cells. It was found that most mutations were C→T transitions at bipyrimidine sites (11,90,91), in fact the mutational hallmark of UV-induced CPDs.

Even if the UVA-induced generation of CPDs is a well-established fact, questions regarding the underlying mechanism of formation have been a matter of debate for a while. CPDs could indeed arise either from a direct photoreaction triggered by the absorption of UVA photons or by photosensitization (Fig. 3). In the latter case, the mechanism may involve excitation of endogenous chromophore(s) with subsequent conversion to long-lived excited triplet states by intersystem crossing. If the energy of the triplet manifolds is larger than that of thymine in DNA, a transfer may take place and lead to the formation of CPDs (92). As will be discussed, later triplet–triplet energy transfer (TTET) process is documented for instance for several phototoxic drugs. Interestingly, chromatographic measurements showed that, in calf thymus DNA triplet transfer photosensitization using ketone compounds leads to the predominant formation of T<>T (90%) together with lower amounts of T<>C and C<>T as observed in UVA-irradiated cells (54,93). An alternative explanation is a direct excitation mechanism that would result from the weak absorption of DNA in the UVA range. It may be added, although being a much lower process than in the UVC and UVB ranges, that absorption of UVA photons by DNA is a well-documented fact (94) that was recently shown to be driven by the stacking of bases in the double helix (95,96). Comparative experiments showed that the yields of UVA-induced CPDs were similar in isolated and cellular DNA (55,96,97). This suggests that endogenous sensitizers play at best a minor role in the formation of these photoproducts. Recent data also shed some light on the origin of the observed differences in the distributions of UVB and UVA bipyrimidine photoproducts. The singlet excited states arising from initial UVA photon absorption that are characterized by a marked charge transfer state appear to be quite different from those induced by UVB irradiation (95). This may explain the lack of 6-4PPs and a different CPD distribution upon exposure of DNA to UVB radiation. Altogether, it seems now most likely that UVA-induced CPDs arise from a direct photoreaction.

Figure 3.

 The two possible mechanisms leading to cyclobutane pyrimidine dimers upon UVA irradiation: (1) direct absorption by DNA bases and (2) photosensitized triplet energy transfer.

UVA-induced oxidation reactions (endogenous photosensitizers)

While bipyrimidine photoproducts are produced by direct excitation of the DNA bases, the formation of oxidatively generated lesions including 8-oxoGua obeys indirect photosensitized mechanisms (Fig. 4). Indeed, neither UVB nor UVA photons provide enough energy to directly enable one-electron oxidation of DNA bases, although the double-stranded structure lowers the ionization potentials of the nucleobases.

Figure 4.

 Different types of oxidatively generated DNA damage induced by singlet oxygen (1O2), photosensitized one-electron oxidation (Phsens*) and hydroxyl radical (?OH).

The diversity of DNA photo-oxidation mechanisms

The UVA-mediated oxidation reactions to cellular DNA have suggested to be initiated by the interaction of incident long UV wavelength radiation with endogenous chromophores (7,22,34). Although this working hypothesis has been under investigation for many years, the exact nature of the involved natural photosensitizers remains unclear (98). Those could be flavin derivatives, porphyrins, specific vitamins and even melanin. The photosensitizers would trigger several DNA oxidative pathways depending on the properties of their excited states and their location in the cell.

Singlet oxygen-mediated oxidation of guanine in DNA.  A common photosensitized pathway involves the production of singlet oxygen (1O2) according to the so-called Type II photosensitization pathway that results from energy transfer from the triplet excited sensitizer to molecular oxygen. O2 is thus converted into a reactive singlet excited 1Δg species that exhibits a strong affinity for molecules rich in double bonds giving rise to dioxetanes, endoperoxides or ene-oxidation products. The unique target among DNA components is guanine that is converted into a 4,8-endoperoxide through 1O2-mediated-Diels Alder [4+2] photocycloaddition reaction. The latter intermediate quantitatively rearranges within ds DNA into 8-hydroxyperoxyguanine (39,99) that is further reduced into 8-oxoGua (Fig. 5), the exclusive 1O2 oxidation product (39,100,101). In early works, it has been suggested that oligonucleotide strand breaks were formed in isolated DNA exposed to 1O2. However, it is now established that 1O2 has no ability to cleave the DNA backbone and that earlier observations concerning formation of DNA nicks would be explained by the occurrence of secondary oxidation of 8-oxoGua that is about 100-fold more reactive than guanine toward 1O2.

Figure 5.

 Photosensitized oxidative degradation pathways of guanine through Type I (one-electron oxidation) and Type II (1O2) mechanisms in cellular DNA.

Photosensitized one-electron oxidation reactions.  A second possible DNA photosensitization pathway involves one-electron oxidation by excited chromophores, also known as the Type I photosensitization mechanism. The efficiency of this reaction is driven by the oxidation potential of the photosensitizers that preferentially act on guanine. Numerous studies have been carried out with isolated bases and nucleosides exposed to UVA or visible light in the presence of specific photosensitizers (7,34). Isolation and characterization of oxidation products have thus been reported for the four DNA bases (103). It was found in DNA that guanine is the main target of one-electron oxidation reactions mediated by Type I photosensitization as the purine base exhibits by far the lowest ionization potential among DNA components (Fig. 5). The three other bases may be oxidized to a minor extent while the sugar phosphate backbone is not a target. The distribution of oxidatively generated damage involving Type I photosensitization has been shown to be modulated, leading to an enhanced guanine specificity by the occurrence of charge transfer within DNA (104,105). DNA damage was revealed by high-resolution polyacrylamide gel electrophoresis as alkali-labile sites, or less frequently by using repair enzymes as cleavage tools. More direct chromatographic analyses led to the same conclusion that guanine is the preferential oxidation target. Upon loss of an electron, guanine is converted into a radical cation (Gua•+) that may undergo hydration with subsequent conversion to reducing the 8-hydroxy-7,8-dihydroguanyl radical. The latter transient radical is subsequently converted to 8-oxoGua and 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FapyGua) by competitive one-electron oxidation and one-electron reduction, respectively. Another reaction of Gua•+ is deprotonation that give rises to a highly oxidizing guanine radical Gua(-H) which after superoxide radical (O2•−) addition leads to 2,2,4-triamino-5(2H)-oxazolone (Oz) through complex reaction pathways including the transient formation of an imidazolone derivative (103).

Hydroxyl radical-mediated DNA damage.  A third oxidative pathway involving the hydroxyl radical (OH) has to be considered in UVA genotoxicity, although not directly initiated by photochemical reactions. OH is a highly reactive oxygen species that plays a major role in radiobiology through indirect effects mediated by water radiolysis. In the case of UVA, OH results from the fate of far less reactive O2•−, that is produced upon Type I photosensitization by reaction of the photosensitizer radical anion with molecular oxygen. O2•− may be also released in large amounts by mitochondria in cells in response to UVA irradiation (106). Subsequently, O2•− can be converted by spontaneous or more likely by superoxide dismutase-driven dismutation into hydrogen peroxide (H2O2). The latter species is the direct precursor of OH upon reaction with reduced transition metal ions such as Fe2+ or Cu+ in a chemical process known as the Fenton reaction. Interestingly, it was shown that UVA leads to the release of iron from ferritin, therefore potentiating emergence of Fenton reactions (107). In contrast to 1O2 or one-electron oxidation, OH does not exhibit target specificity and reacts with all components of DNA at diffusion-controlled rates (108). The reaction of OH with purine bases leads to the transient formation of the C8-hydroxylated radical that gives rise to 8-oxoGua and FapyGua together with related adenine degradation products including 8-oxo-7,8-dihydroadenine (8-oxoAde) and 4,6-diamino-5-formamidopyrimidine (FapyAde), however, in much lower yields. It is also well documented that OH is able to react with pyrimidine bases at the C5–C6 double bond, leading to the formation of 5,6-dihydroxy-5,6-dihydrothymine (ThyGl) and 5,6-dihydroxy-5,6-dihydrocytosine (109). 5-Hydroxymethyluracil (HmUra) and 5-formyluracil (ForUra) have also been characterized as the main OH-mediated methyl oxidation products of thymine (109). It may be added that OH efficiently reacts with the 2-deoxyribose moieties by hydrogen abstraction according to several pathways that lead in most cases to the formation of DNA strand breaks (109).

UVA-mediated oxidatively generated DNA damage in cells

A number of studies have been devoted to the identification and quantification of DNA oxidation products in UVA-irradiated cells. DNA–protein crosslinks (DPCs) were among the first lesions suggested to be generated upon UVA irradiation. However, no clear structural identification of the DPCs was made (110). A possible mechanism that is discussed later would involve addition of nucleophilic amino acids like lysine to the guanine radical cation as recently observed in model systems (111). A much more common biomarker of oxidatively generated damage to DNA is 8-oxoGua that has been found to be generated in significant yields upon exposure of cultured cells (85,112,113) and skin (61) to UVA radiation. This was achieved using HPLC associated with either electrochemical or tandem mass spectrometry detection. As mentioned above, 8-oxoGua that is a ubiquitous biomarker of oxidative DNA degradation reactions may be produced by 1O2, one-electron oxidation and OH. Therefore, detection of 8-oxoGua that is strongly indicative of occurrence of oxidative degradation reactions in DNA is however poorly informative in terms of the mechanisms involved. In contrast, observation of the formation of strand breaks in DNA provides strong support for the implication of OH. It was shown using the comet assay and the alkaline elution analysis that UVA irradiation of cells gave rise to the formation of alkali-labile sites, mostly strand breaks. It was recently confirmed that UVA radiation, in contrast to earlier suggestions, is not able to generate double strand breaks (114).

Most of the techniques designed for the quantification of strand breaks may also provide information on the formation of modified bases upon their conversion into additional breaks by using purified repair enzymes such as DNA N-glycosylases. The most commonly used proteins include bacterial formamidopyrimidine glycosylase (Fpg) and endonuclease III (Endo III) that recognize oxidized purine and pyrimidine bases, respectively. It was found that, oxidized purine bases and most likely 8-oxoGua, are the main UVA-induced oxidation products in UVA-irradiated cells (32,54,115). The second most frequent UVA-induced DNA oxidatively generated lesions are strand breaks that are formed in an approximately three-time lower yield than oxidized purines. The level of oxidized pyrimidines is less than half of that of strand breaks. Comparison with the molecular effects of other oxidative stress agents including ionizing radiation and 1O2 has provided useful mechanistic insights into the UVA-mediated oxidation of cellular DNA. Exposure of cells to low linear energy transfer ionizing radiation like X- or γ-rays that act mostly through the production of OH gives rise to strand breaks and oxidized bases in a 1.3:1 ratio while the formation yields of purine and pyrimidine oxidation products are similar. A very different situation is noted when specific exogenous Type II photosensitizers that lead to the sole formation of 8-oxoGua are used (115). It was thus concluded that UVA formation of oxidatively generated damage to cellular DNA was mostly rationalized in terms of predominant implication of 1O2 with a minor contribution of OH (25% in the case of 8-oxoGua). It should be kept in mind that the distribution of UVA-induced DNA oxidation products may depend on the wavelength of the incident UVA photon as illustrated by the action spectra for the formation of the different classes of lesions (30,32). Another important parameter that may affect the distribution of UVA-mediated DNA degradation products is cell type. This was recently illustrated for melanocytes that are more susceptible than other cell types to UVA-induced oxidatively generated damage to DNA both as isolated cells (89,116) and in pigmented mice (117). A summary of the main information available on the distribution of DNA damage generated upon exposure of cells to UVB and UVA radiations is provided in Fig. 6.

Figure 6.

 Typical relative distribution of the main types of DNA lesions in UVB- and UVA-irradiated cells.

Ionization of nucleobases by high-intensity UVC laser pulses

Exposure to laser beams has been used as an efficient tool to damage DNA. For instance, infrared (IR) and UVA laser sources have been applied to induce DNA modifications in cells and investigate subsequent cellular responses such as repair (118–122). Although studies on isolated DNA have shown that three-photon absorption of IR gave rise to CPDs (123), the formation of several other lesions was mostly explained by the production of reactive oxygen species. UV laser irradiation has been performed to study the effect of base ionization within DNA. The first studies focused on the delineation of modifying effects of vacuum UV (193 nm) on plasmid DNA (124–126). It was found that, like for Type I photosensitization, guanine was the most frequently damaged base, although cytosine, thymine and adenine were initially ionized upon exposure to laser pulses. It may be noted that direct strand breakage was a minor process. Similar observations were made following exposure of isolated genomic DNA exposed to high-intensity UVC pulsed radiation that operate on nucleobases through biphotonic ionization (127). Absorption of the first photon by a nucleobase gives rise to an excited intermediate that is able to absorb a second photon due the high intensity of the nanosecond laser pulses. The level of energy thus absorbed is larger than the ionization threshold of the base and as a result the corresponding radical cation is generated. This reaction that is pulse intensity dependent occurs concomitantly with the depletion of the pyrimidine bases in the single excited states. Therefore, an increase in laser pulse intensity is accompanied with a decrease in the quantum yield of CPDs and 6-4PPs. In contrast, the level of 8-oxoGua, the main DNA oxidation product, and of other oxidized bases including 8-oxoAde, ThyGl, 5-HmUra and 5-FoUra increases (128). This clearly shows that biphotonic ionization takes place and that, like for other reactions leading to the formation of radical cations in DNA, guanine is the main direct or indirect target of one-electron oxidation. The formation of 8-oxoGua, 8-oxoAde and ThyGl is rationalized in terms of initial hydration reaction of the guanine, adenine and thymine radical cations, respectively, whereas the generation of HmUra and FoUra involves initial deprotonation of thymine radical cation (129,130). This does not only apply to isolated DNA as similar results were obtained in human cells exposed to high-intensity nanosecond UVC pulses (128). This represents one of the few direct observations of the actual occurrence of hole migration within cellular DNA.

Sensitized reactions to UVA radiation by exogenous compounds

As already mentioned, DNA only absorbs slightly in the lower part of the UVA range of solar radiation up to 340 nm, leading mostly to the formation of cis–syn T<>T through delocalization of excited state transients on several nucleobases (95,96). Evidence has been provided for the occurrence of several UVA-sensitized reactions in both isolated and cellular DNA that are triggered by exogenous compounds, mostly drugs that are used for therapeutic purposes (7,131–133). We may distinguish photosensitizers which operate in an oxygen-independent manner through mainly triplet energy transfer and photoaddition reactions. A second class of photosensitized reactions requires the presence of oxygen in order to promote damage to target biomolecules according to the so-called photodynamic effects in cells. These include Type I and Type II photosensitization mechanisms that involve predominantly one-electron transfer (or hydrogen atom abstraction) mediated by triplet excited photosensitizers and the formation of singlet oxygen by energy transfer, respectively (7,103,134). In that respect guanine (135) is the major target of one-electron oxidation reactions of most Type I photosensitizers (136). A few photosensitizers including anthraquinone (137), benzophenone (138), 1,4-dimethyl-2-naphthoquinone (139) and riboflavin (140) are able to abstract one electron from adenine and pyrimidine bases in addition to guanine. However, the final DNA degradation products are predominantly generated at guanine sites as the result of hole transfer migration and redistribution of initial radical base damage (104,141). Several UVA-sensitized reactions to cellular DNA by exogenous agents have been identified through the isolation and characterization of specific base degradation products as further discussed below.

Triplet energy transfer

There is an abundant literature on the TTET-mediated formation of CPDs in ds DNA by several photosensitizers, including acetone (142), acetophenone (54,142), benzophenone (54,143), carprofen (144), fenofibric acid (143), fluoroquinolones (92,93,145), ketoprofen (143) and three pyridopsoralen derivatives (146,147). The predominant CPD photoproduct of the latter photosensitized reactions was found to be cis–syn T<>T, in agreement with the fact that thymine exhibits the lowest triplet energy (ET) among nucleobases (148). It was recently found using several fluoroquinolones exhibiting ET values within the energy range 253–273 kJ mol−1 that triplet energy of thymine in ds DNA was 267 kJ mol−1 (113). The decrease by about 43 kJ mol−1 of the thymine ET value with respect to that of the isolated base was suggested to result from the occurrence of π-stacking and base pairing in DNA (92). Several examples of UVA-sensitized formation of CPDs by fluoroquinolones in the DNA of isolated cells (93,149,150) and mouse skin (151) are available. Thus, UVA-excited lomefloxacin has been found to generate in human keratinocytes CPDs that were revealed as T4 endonuclease V-sensitive sites (149). CPDs were also detected by immunoassay in the DNA of fibroblasts, keratinocytes and Caucasian melanocytes upon lomefloxacin photosensitization (150). Evidence has been also provided using an immunohistochemical detection approach for the lomefloxacin-sensitized formation of CPDs in the skin of mice XPA+/+ and XPA−/− upon UVA irradiation (151). Further detailed information was gained on the photosensitizing properties of several fluoroquinolones on the DNA of THP-1 human monocytes (93). It was found that in most cases TTET was the predominant mechanism of photosensitized damage to DNA over the formation of 8-oxodGuo and DNA strand breaks. The efficiency of formation of T<>T that was assessed using the accurate HPLC-ESI-MS/MS method was found to decrease in the following order: enoxacin > norfloxacin > lomefloxacin (93). Carprofen, a nonsteroidal anti-inflammatory drug, is also able to UVA sensitize the formation of CPDs in the DNA of human keratinocyte (HaCaT) cell lines as inferred from immunofluorescence detection and application of a modified version of the alkaline comet assay involving the use of T4 endonuclease V (152). It may be emphasized that the photosensitized formation of CPDs by fluoroquinolones and carprofen in either cells or skin implies that the photosensitizers are located in the close vicinity of nuclear DNA in order to allow efficient TETT.

Photocyloaddition reactions of psoralen derivatives

Furocoumarins including bifunctional 8-methoxypsoralen (8-MOP) and 5-methoxypsoralen (5-MOP) are still used in the so-called “psoralen plus UVA (PUVA)” therapy of rather frequent hyperproliferative skin diseases, including psoriasis and vitiligo (153–155). Other applications of psoralen derivatives deal with the photopheresis of cutaneous T-cell lymphoma (156–158) and the photoinactivation of pathogens in plasma and platelet components (159,160). Major mechanistic insights into the mechanisms of [2+2] photocycloaddition of furocoumarin derivatives to the 5,6-double bond of thymidine, the major target of these oxygen-independent photoreactions, have been gained from extensive model studies (35,161–163). Essentially, the formation of two types of monoadducts as pairs of diastereomers through the implication of either the 4′,5′-double bond of the furan moiety or the 3,4-ethylenic bond of the pyrone ring has been reported. While pyrone-side monoadducts to thymine are not photoreactive, UVA excitation of furan-side monoadducts may lead to the formation of harmful DNA interstrand cross-links (ICLs) through cycloaddition to opposite thymine at 5′-TpA-3′ sequences (Fig. 7) that are the preferential sites of photoreactions of most psoralen derivatives (35,161–163). There is an abundant literature on the indirect methods of measurement of monoadducts and ICLs based mostly on the use of alkaline elution, enzyme-linked and immunostaining assays (for a review on earlier reported methods, see Cadet et al. [35]). Recently, a modified version of the alkaline comet assay that includes a postlysis γ-irradiation step with a dose of 9 Gy has been designed for measuring DNA ICLs induced in HaCat cells upon treatment with 8-MOP plus UVA radiation (164). This appears to be the most sensitive method available so far for monitoring the formation of psoralen-DNA biadducts and assessing their subsequent repair. The repair rate of ICLs that is rather slow, requiring at least 24 h to be completed, was found to decrease with an increase in the concentration of 8-MOP (164). The first individual measurement of biadducts and furan- and pyrone-side mono adducts formed by photoreaction of 8-MOP with thymine in the DNA of Saccharomyces cerevisiae yeast cells has been achieved using a lipophilic Sephadex LH-20 column (165). This has required the use of radiolabeled 8-MOP for a sensitive detection and the release of adducts as modified nucleobases by acidic hydrolysis treatment. Almost simultaneously, a HPLC fluorescence method has been reported for the measurement of the two cis–syn diastereomers of the furan-side monoadducts of 3-carbethoxypsoralen (3-CPs) to thymidine in the DNA of S. cerevisiae and Chinese hamster V79 cells treated with monofunctional 3-CPs plus UVA radiation (166). A mild enzymatic digestion of DNA has been performed to ensure a quantitative release of the modified nucleosides. The method is sensitive enough to detect each of the two diastereomers that are formed with a yield of about 1 lesion per 104 base pairs in cellular DNA (166). This has also allowed the determination of the kinetics of removal of the two 3-CPs furan-side monoadducts to thymidine in yeast cells (167). The fast disappearance in the postincubation repair period of the two diastereomers with a half-life of about 90 min would suggest implication of the base excision repair pathway for their enzymatic removal. It may be added that the two cis–syn diastereomers of 5-MOP to the furan-side to thymidine has been also measured in the DNA of S. cerevisiae using the same approach (168). However, one of the main limitations of the fluorescence detection is that only furan-side monoadducts to pyrimidine bases can be measured. This has been recently overcome by the use of the versatile, specific and accurate electrospray ionization tandem mass spectrometry (ESI-MS/MS) technique in association with HPLC following nuclease P1 digestion of DNA (169). Therefore, ICLs induced by UVA activation of 8-MOP in the DNA of human skin melanoma cells were detected as tetranucleotides with a threshold sensitivity close to 1 ICL per 106 normal nucleosides. Application of the method has been then extended to the measurement of furan- and pyrone-side monoadducts of 8-MOP and 3-(2-amino-ethoxymethyl)-2,5,9-trimethyl-furo[3,2-g]chromen-7-amotosalen (S59) to thymidine with, however, only a tentative assignment of the two types of monoadducts released after nuclease P1 digestion as dinucleotides (170). Standard ICL and monoadducts of D3-8-MOP-containing oligonucleotides were used for the calibration of the HPLC-MS/MS assay. It was found that furan- and pyrone-side monoadducts of both 8-MOP and S59, a recently developed psoralen derivative for pathogen inactivation in blood products (171,172), are formed upon UVA irradiation in a much lower yield (close to 1%) than ICLs. A much higher photoreactivity of S59, up to 2 orders of magnitude, with respect to 8-MOP was inferred from the measured levels of mono-adducts and ICLs. A complementary analytical tool for monitoring the distribution of psoralen mono- and bi-adducts to pyrimidine bases in duplex oligodeoxynucleotides that involves the association of liquid chromatography with either UV/ESI-MS or IR multiphoton dissociation mass spectrometry has recently become available (173).

Figure 7.

 [2+2] Photocycloaddition reactions of 8-methoxypsoralen giving rise to interstrand cross-links and furan- and pyrone-side monoadducts upon UVA irradiation.

Table 1.   Yield* of formation of bipyrimidine photoproducts in primary cultured of human keratinocytes and in human skin exposed to either UVB or UVA (data from Mouret et al. [61]).
 T<>T6-4 TTT<>C6-4 TCC<>TC<>C
  1. n.d. = not detected. *Reported data represent the yield ± standard deviation for six donors.

UVB (yield in photoproducts/104 bases J−1 cm−2)
 Skin2.06 ± 0.710.19 ± 0.151.58 ± 0.260.72 ± 0.100.56 ± 0.080.44 ± 0.24
 Keratinocytes45.23 ± 6.453.33 ± 0.4929.30 ± 6.3618.00 ± 2.4811.86 ± 2.506.01 ± 1.40
UVA (yield in photoproducts/104 bases kJ −1cm−2)
 Skin0.69 ± 0.21n.d.0.12 ± 0.06n.d.0.06 ± 0.04n.d.
 Keratinocytes1.14 ± 0.08n.d.0.08 ± 0.05n.d.0.07 ± 0.02n.d.

Oxidation reactions mediated by photosensitizers

Numerous compounds are able to oxidize DNA according to the Type I and/or Type II photosensitization mechanisms. It is well documented that the reaction of 1O2 (Type II ROS) leads to the specific formation of 8-oxoGua whereas Type I gives rise, among other degradation products, 8-oxoGua and 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FapyGua), to the transient formation of the guanine radical cation (Gua•+) in DNA. In that respect, several detailed mechanistic studies have provided relevant information on the contribution of the oxidation mechanisms to the degradation of nucleosides and isolated DNA (174–176). Among fluoroquinolones (for recent reviews, see Lhiaubet-Vallet et al. [175] and de Guidi et al. [176]), lomefloxacin and, to a lesser extent, enoxacin act as predominant Type I photosensitizers whereas ofloxacin and norfloxacin react with DNA mainly through 1O2 oxidation (93). It was subsequently confirmed that ofloxacin oxidized predominantly DNA mostly through the Type II mechanism (177) as also shown for norfloxacin (178). Formation of 8-oxoGua was observed in the DNA of cells photosensitized by norfloxacin (93) and rufloxacin (179,180). Further support for the predominant formation of 8-oxoGua among the DNA damage induced by UVA-excited rufloxacin UVA was provided by the enhanced frequency of GC>TA transversions in yeast after fluoroquinolone photosensitization (181). It was not possible, however, in the above photochemical studies to assess the relative contribution of Type I and Type II mechanisms of DNA oxidation, likely to mostly involve 1O2, as no specific one-electron oxidation decomposition guanine products such as FapyGua or Oz were measured. The same remark applies to the detection of 8-oxoGua and Fpg-sensitive sites in cellular DNA upon UVA sensitization by methylene blue (182) and riboflavin (183), respectively.

Fluorescent light-excited zinc meta N-methylpyridylporphyrin has been shown to induce oxidatively generated damage to DNA of human colon carcinoma LS174T cells, tentatively assigned as 8-oxoGua (184) on the basis of somewhat questionable immunodetection (185). Cationic 5,10,15,20-tetrakis (N-methyl-4-pyridyl)-21H,23H-porphyrin that is mostly localized in the nuclei of human leukemia HL-60 cells as inferred from confocal fluorescence microscopy is able to UVA sensitize the formation of 8-oxoGua through predominant Type II photosensitization mechanisms (186). In contrast, protoporphyrin IX derived from 5-aminolevulinic acid and whose preferential localization was observed in the cytoplasm was found to generate only small amounts of 8-oxoGua upon UVA illumination (186).

Nanoparticles (ca 100 nm or smaller) of titanium dioxide (TiO2), a white pigment, are widely used for various industrial purposes, including the design of sun blockers and other cosmetic products (187,188). UVA-excited TiO2 has been shown to generate 1O2 (189), O2 and also highly reactive OH either through reduction of surface hydroxyl group by valence bond holes (190) or by Fenton type reactions (191). Photoexcited TiO2 exhibits strong bactericidal activity (191–194) and genotoxicity (195–198) on various cells as mostly the result of oxidative stress. Sunlight illumination of MRC-5 fibroblasts preincubated with TiO2 was found to generate DNA strand breaks and/or alkali-labile sites as inferred from alkaline single-cell gel electrophoresis (199). Using alkaline comet analysis the anatase form of TiO2 was shown to exhibit a higher photocatalytic ability than rutile to induce DNA damage in human bronchial epithelial cells (200). Support for the implication of OH in the UVA activation of nanocrystalline TiO2 was provided by atomic force microscopy (AFM) measurements showing a marked decrease in the stiffness of phototreated human skin fibroblasts (201). OH has been also proposed to explain the increased level of oxidatively generated damage to DNA measured as Fpg-sensitive sites in the DNA of UVA-irradiated goldfish skin GFSk-S1 cells pretreated with TiO2 nanoparticles (202). However, the lack of increase in the level of strand breaks and Endo III-sensitive sites would argue in favor of a significant contribution of 1O2 in the TiO2-photosensitized generation of oxidatively generated damage rather than of OH. One may also quote a recent study in which preincubation of S. cerevisiae cells with normal size TiO2 (>100 nm) led to an increase in mutation frequencies in both wild-type and 8-oxoguanine DNA N-glycosylase strains upon UVB irradiation (203). This that was partly prevented by the presence of metal chelators such as dipyridyl and neocuprine is suggestive of the predominant role of OH in TiO2-sensitized formation of DNA damage to UVB. In light of the above preliminary observations that cannot provide any definitive conclusions, there is a strong need of further studies in order to better assess the mechanisms involved in the TiO2 photosensitization of cellular DNA.

Oxidatively generated and complex DNA lesions induced by UVA activation of thiopurines

Thiopurines including azathioprine, 6-mercaptopurine and 6-thioguanine (6-TG) are widely used in cancer therapy, for the treatment of inflammatory disorders and as immunosuppressants (204,205). However, a significant increase in the incidence of skin squamous carcinoma has been shown to be associated with long-term use of thiopurines in relation to solar radiation exposure, particularly in transplant patient recipients (204,206). An indication for the induction of the deleterious effects triggered by UVA activation of thiopurines was provided indirectly by the observed inhibition of transcription (207) and abrogation of cell-cycle checkpoints (208) in cultured human cells treated with thiopurine and UVA radiation through the generation of DNA damage. This is explained by the metabolization and incorporation of thiopurines into DNA as 6-thio-2′-deoxyguanosine, an efficient endogenous Type I and Type II photosensitizer absorbing in the UVA range with a maximum at 342 nm (209). The steady-state photophysics and excited-state dynamics of 6-TG in aqueous solutions have been recently investigated (210). In agreement with quantum chemical calculations, it was found that the singlet excited state of the base was converted efficiently by very fast intersystem crossing into the triplet manifold. The high triplet quantum yield value (ΦT = 0.8 ± 0.2) is likely to favor energy transfer to O2 and therefore the generation of 1O2 (210). The quantum yields of 1O2 generation by UVA-excited 6-TG and related ribonucleosides that were almost simultaneously determined by time-resolution luminescence measurements at 1270 nm were found to vary within the 0.40–0.58 range (211). The main DNA oxidation products arise from 1O2 oxidation of the thiol function of 6-TG; they were characterized as stable guanine 6-sulphinate (GS02) and guanine 6-sulfonate (GS03) through initial transient guanine 6-sulphenate (GSO) (212,213). However, the formation yields of stable single oxidation products, including GS02, GS03 and 8-oxoGua (206) in 6-TG containing isolated and cellular DNA are not available so far. The formation of two other main classes of 6-TG-photosensitized photoproducts, including DPCs and DNA ICLs, was also observed in model compounds and CCRF-CEM leukemia cells (214,215). Some of the proposed structures of DPCs formed between oligonucleotides and short peptides involve covalent binding of highly reactive 6-TG with cysteine residues through the generation of a disulfide (214). Free peptide amino groups were also suggested to participate in the 6-TG-mediated formation of DPCs. However, the DPCs that were generated in cellular DNA were resistant to reduction treatment in contrast to those formed in model studies (214), ruling out the possible structures that were discussed above. An alternative mechanism for the 6-TG-photosensitized formation of DPCs has been recently proposed (216) on the basis of the well-documented reactivity of Gua•+, the predominant one-electron DNA oxidation transient radical formed, with relevant biological nucleophiles (111,134,136,217). In addition to water that was found initially to efficiently react with Gua•+ (218), several amino acids bearing a nucleophilic group, including lysine, arginine and serine were shown to covalent bind to the C8 of guanine (111,216). A detailed model study of the formation of DPC has shown that the central lysine of KKK peptide was able to undergo highly efficient covalent attachment through its free ε-amino group to one-electron oxidation guanine intermediate of the TGT trinucleotide (111) (Fig. 5). Other independent proofs that support the covalent attachment of nucleophilic amino acids to Gua•+ upon one-electron oxidation of oligonucleotides mediated by a Type I photosensitizer, the flash quench technique or biphotonic ionization, are available (219–222). Similarly, the 6-TG photosensitized formation of ICLs in cellular DNA could be also accounted for the covalent bond formation between opposite cytosine to Gua•+ as observed upon exposure DNA duplexes to high-intensity UVC nanosecond pulses. However, further studies including isolation and characterization of DPCs and ICLs induced by photoactivation of incorporated 6-TG in cellular DNA are required to further ascertain the proposed mechanism involving one-electron oxidation of guanine. Evidence was recently provided for the incorporation of 6-TG in mitochondrial DNA of HCT116 human cells (224). This was found to lead to the generation of ROS upon UVA irradiation with subsequent formation of DNA oxidation products that is accompanied by a loss of mitochondrial function (224).

4-Thiothymine (S4Thy), a thymine analog absorbing in the UVA range with a maximum centered around 340 nm is able to efficiently incorporate into the DNA of dividing cells (225, 226). This leads to a high phototoxicity of the cells upon UVA-irradiation. This was rationalized in terms of oxygen-independent formation of DNA photoproducts including an expected S5 thietane derivative, a 6-4PP analogue (227), through photocycloaddition with vicinal thymine and putative ICLs (228) that however remain to be characterized. Attempts to search for an increase in the level of 8-oxoGua in cellular S4Thy DNA were unsuccessful despite the fact that S4Thy is an efficient generator of 1O2 upon UVA irradiation (229).

Conclusions

Major progress in the measurement of DNA photoproducts in isolated cells and human skin exposed to either UVB, UVA and simulated solar radiation has been accomplished during the last decade. This is complemented by a better understanding of the photochemical pathways involved in the formation of DNA photoproducts in irradiated cells, mostly as the result of the availability of comprehensive mechanisms from detailed model studies. In that respect it remains to further explore the possibility of generation of other UVB-induced DNA lesions including TA photoproducts (44) and intra-strand C(4-8)G and G(8-4)C photoproducts (230). There is also an increasing interest in delineating photosensitized mechanisms triggered by drugs including fluoroquinolones and thiopurines. In particular it remains to further identify the mechanisms of formation of DPCs and ICLs that are induced by UVA excited 6-thioguanine once incorporated in cellular DNA.

Appendix

Author biographies

inline image

Jean Cadet is currently Scientific Adviser at CEA/Grenoble and Adjunct Professor at University of Sherbrooke. He received his Ph.D. in chemistry from the University of Grenoble in 1973 and has been the Head of Laboratory of Lõsions des Acides Nuclõiques at the French Atomic Energy Commission, CEA/Grenoble until 2001. His research interests include various aspects of the chemistry and biochemistry of oxidatively generated and photo-induced damage to DNA (mechanisms of reactions, measurement in cells, assessment of biological features including substrate specificity of DNA repair enzymes and mutagenesis of base lesions). He has received several awards including Research Award from American Society for Photobiology, the medal of Excellence from European Society for Photobiology, the Charles Dhõrõ Award and Berthelot Medal from the French Academy of Sciences.

inline image

Stéphane Mouret has devoted his entire research activity to skin biology. During its PhD work at Grenoble Hospital in the Department of Dermatology, he studied the relationship between human papillomavirus infection and UV radiation in skin carcinogenesis and was graduated from the Claude Bernard University, Lyon, France, in 2003. In 2005, he changed group and began a post-doctoral work at the French Atomic Energy Commission, Grenoble, France, and his research was focused in the field of photobiology and photoprotection in cells and human skin, with emphasis on UV-induced DNA damage and particularly the genotoxicity of UVA. He is author of numerous scientific articles concerning photobiology of DNA. Presently, he is researcher at Armed Forces Biomedical Research Institute, La Tronche, France and studied the impact of chemical toxic agents on skin to understand the physiopathology of skin lesions and to develop therapeutic treatments.

inline image

Jean-Luc Ravanat graduated in Biochemistry and Organic Chemistry and focused his work on the identification of the nature and mechanisms of formation of DNA lesions. He initially devoted efforts in understanding photosensitized reactions of nucleic acids including the role of singlet oxygen. Now he is more involved in studying the effect of ionizing radiation, in particular regarding the mechanisms of formation of radiation-induced complex DNA lesions. In addition he has made a significant contribution in the development of sensitive and accurate assays to measure DNA lesions at the cellular level.

inline image

Thierry Douki is a scientist at the Atomic Energy Commission in Grenoble, France. He was trained in organic synthesis. Most of his research activities focus on DNA damage produced by a wide variety of genotoxic agents including oxidative stress, chemicals and ionizing and ultraviolet radiation. In this last field, he is mostly involved in the characterization of photoproducts, the development of analytical tools for their quantification and the study of the formation and repair of UV-induced DNA damage in cells and skin. Thierry Douki is Associate-Editor of Photochemistry and Photobiology, member of the European and American Societies for Photobiology. He is also secretary of the French Society for Photobiology. He is co-author of more than 130 original articles and a number of reviews and book chapters on various aspects of DNA damage.

Ancillary