**2. Mechanism of action**

The potential target of metallic elements on DNA repair proteins are the zinc finger structures in their DNA binding motifs. Within these structures, zinc is complexed to four cysteines and/or histidines, folding different structural domains mediating DNA-protein as well as protein-protein interactions. It is estimated that about 1% of all mammalian genes encode zinc finger proteins, which are involved in many processes maintaining genomic integrity (Mackay & Crossley, 1998). The zinc ions do not participate in interactions conveyed by zinc finger domains, but are necessary for their function since they maintain their three-dimensional structures. In the case of transcription factors and DNA repair proteins, the absence of metal ions lead to loss in DNA-binding capacity. The functions of individual zinc finger include recognition of structures and sequences of nucleic acids and proteins. The majority of identified zinc finger may be classified as transcription factors. Another well known function of various zinc finger motifs, is the assembly of multiprotein

Interactions by Carcinogenic Metal Compounds with DNA Repair Processes 31

proteins. The metal ions can oxidize the essential cysteines and/or other residues in zinc finger structures interfering in metal binding domain. Taken together, the above mentioned mechanisms indicate that DNA repair, zinc homeostasis, oxidative assault and the redox status of the cell are all interconnected (Fig 1). The toxic/carcinogenic metals with sufficiently high affinities to thiols may interfere at all stages of zinc homeostasis and signaling, but specific ways of their actions can only be understood in appropriately complicated experimental designs. Yet each zinc finger protein exerts its own structural function toward metallic compounds but no general prediction about this phenomenon

Fig. 1. Schematic representation of potential interactions of metallic elements with zincbinding structures in transcription factors and DNA repair protein (modified from Hartwig

Different mechanism of action have been suggested for arsenic carcinogenicity including the induction of oxidative stress, induction of genetic damage, altered DNA methylation patterns, enhanced cell proliferation, inhibition of the tumor suppressor protein p53, DNA repair alteration and recently biomethylation (Aposhian & Aposhian 2006). A possible molecular mechanism for arsenic toxicity may lie in its ability to react with thiols, for example, in zinc binding structures of transcription factors, cell cycle control and DNA repair proteins (Kitchin & Wallace, 2008). Nucleotide excision repair (NER) in particular is strongly inhibited by arsenic. NER is capable of removing a wide variety of bulky, DNA helix distorting lesions, caused, e.g., by UV-irradiation or environmental mutagens. Arsenic is known to enhance the persistence of bulky DNA lesions and consequently the mutagenicity induced by UV and benzo[a]pyrene (Hartmann & Speit 1996; Gebel, 2001). Since bulky lesion formation is the possible responsible for their carcinogenicity, genetic integrity depends largely on NER efficiency. Many studies have shown that inorganic arsenic inhibits repair of bulky DNA adducts induced by UV-irradiation (Hartwig et al., 1997; Danaee et al., 2004) or benzo[a]pyrene in cultured cells and laboratory animals (Schwerdtle et al., 2003; Shen et al., 2008); additionally arsenite has been shown to downregulate expression of some NER genes in cultured human cells (Hamadeh et al., 2005). In

appear to be possible.

et al.,2001).

**3. Arsenic** 

complexes having structural or enzymatic functions. The main zinc finger proteins involved are the bacterial formamidopyrimidine-DNA glicosylase (fpg), the xeroderma pigmentosum A (XPA) protein, the poly(ADP-ribose) polymerase (PARP) and tumor suppression protein p53. The studies about of metal ions interaction indicated that apparently similar zinc finger domains may have different reactivity and suggested to draw a sort of list of possible mechanisms of zinc finger damage (Hartwig, 2001). Among these mechanisms, the most important are isostructural substitution, substitution with altered geometry, mixed complex formation, and catalysis of thiol oxidation. Many studies showed that different concentrations are needed to observe an inhibition and wide differences were observed for example when comparing the results obtained with Fpg and XPA (Witkiewicz-Kucharczyk & Bal., 2006). Regarding XPA, arsenic and lead did not decrease its binding to a UVirradiated oligoucleotide, whereas cadmium, cobalt and nickel interfere with its DNA binding ability. A simultaneous treatment with zinc largely prevented this inhibition (Asmuß et al., 2000). Structural investigations by different spectroscopic methods revealed a tetrahedrally co-ordination of all three metal ions with no major distortion of XPA while for cadmium an increased Cd-S bond length was observed. Poly(ADP-ribosyl)ation of various proteins is one of the earliest nuclear events following DNA strand break induction. Yager & Wiencke (1997) and Hartwig et al., (2002) demonstrated an inhibition of H2O2-induced PARP activity in intact cells by nickel, cobalt, cadmium and very low concentration of arsenic in HeLa cells, while no effect was observed with lead. One other zinc-dependent protein with great impact on the processing of DNA damage and genomic stability is the p53 suppressor protein. Zinc has been shown to be required for proper folding in wild type conformation, and exposure to either isolate p53 protein or human breast cancer cells to cadmium resulted in disruption of native p53 conformation and inhibition of DNA binding (Meplan et al., 1999). Witkiewicz-Kucharczyk & Bal (2006) assessed the different metal binding properties of zinc finger and reported that cobalt is practically isostructural with zinc in zinc finger peptides and proteins, regardless of the number of cysteine residues involved. However, there is a strong, although quantitative variable thermodynamic preference for zinc over cobalt in tetrahedral environments provided by zinc finger. This effect is due to the ligand field stabilization effect, modulated by entropic contributions (Lachenmann et al., 2004). The substitution of nickel into zinc finger can also be achieved, but it results in distortions of the binding geometry and alterations of the peptide fold. A distorted tetrahedral coordination was found for cysteine2-Histidine2 zinc fingers (Posewitz & Wilcox, 1995), while a nearly square planar arrangement of donors was demonstrated in a cysteine4 environment (Bal et al., 2003). Also the efficacy of cadmium binding is related to the number of coordinated thiolates (Krizek et al., 1993). Zinc seems to be preferred in the cysteine2-histidine2 environment, the affinities of zinc and cadmium may be comparable for cysteine3-histidine, and cadmium is strongly preferred in cysteine4 zinc finger peptides (Kopera et al., 2004). This preference is due to the high enthalpy of the cadmium-S bond. The cadmium ion fits into tetrahedral environments with little strain, however, it is significantly larger than zinc and cobalt ions, which results in local distortions of zinc finger geometries (Buchko et al., 2000). Finally lead, element with high affinities to thiolates, can replace zinc in zinc finger domains and disrupt their fold. Arsenic too is known to have a high affinity to -SH groups and it is demonstrated by arsenic-mediated repair inhibition not related with a direct mechanism of one or more specific repair proteins, but rather with changes in gene expression and/or signal transduction (Asmuß et al., 2000). Finally redox regulation has been *in vitro* and *in vivo* demonstrated in several DNA-binding zinc finger proteins. The metal ions can oxidize the essential cysteines and/or other residues in zinc finger structures interfering in metal binding domain. Taken together, the above mentioned mechanisms indicate that DNA repair, zinc homeostasis, oxidative assault and the redox status of the cell are all interconnected (Fig 1). The toxic/carcinogenic metals with sufficiently high affinities to thiols may interfere at all stages of zinc homeostasis and signaling, but specific ways of their actions can only be understood in appropriately complicated experimental designs. Yet each zinc finger protein exerts its own structural function toward metallic compounds but no general prediction about this phenomenon appear to be possible.

Fig. 1. Schematic representation of potential interactions of metallic elements with zincbinding structures in transcription factors and DNA repair protein (modified from Hartwig et al.,2001).

#### **3. Arsenic**

30 Selected Topics in DNA Repair

complexes having structural or enzymatic functions. The main zinc finger proteins involved are the bacterial formamidopyrimidine-DNA glicosylase (fpg), the xeroderma pigmentosum A (XPA) protein, the poly(ADP-ribose) polymerase (PARP) and tumor suppression protein p53. The studies about of metal ions interaction indicated that apparently similar zinc finger domains may have different reactivity and suggested to draw a sort of list of possible mechanisms of zinc finger damage (Hartwig, 2001). Among these mechanisms, the most important are isostructural substitution, substitution with altered geometry, mixed complex formation, and catalysis of thiol oxidation. Many studies showed that different concentrations are needed to observe an inhibition and wide differences were observed for example when comparing the results obtained with Fpg and XPA (Witkiewicz-Kucharczyk & Bal., 2006). Regarding XPA, arsenic and lead did not decrease its binding to a UVirradiated oligoucleotide, whereas cadmium, cobalt and nickel interfere with its DNA binding ability. A simultaneous treatment with zinc largely prevented this inhibition (Asmuß et al., 2000). Structural investigations by different spectroscopic methods revealed a tetrahedrally co-ordination of all three metal ions with no major distortion of XPA while for cadmium an increased Cd-S bond length was observed. Poly(ADP-ribosyl)ation of various proteins is one of the earliest nuclear events following DNA strand break induction. Yager & Wiencke (1997) and Hartwig et al., (2002) demonstrated an inhibition of H2O2-induced PARP activity in intact cells by nickel, cobalt, cadmium and very low concentration of arsenic in HeLa cells, while no effect was observed with lead. One other zinc-dependent protein with great impact on the processing of DNA damage and genomic stability is the p53 suppressor protein. Zinc has been shown to be required for proper folding in wild type conformation, and exposure to either isolate p53 protein or human breast cancer cells to cadmium resulted in disruption of native p53 conformation and inhibition of DNA binding (Meplan et al., 1999). Witkiewicz-Kucharczyk & Bal (2006) assessed the different metal binding properties of zinc finger and reported that cobalt is practically isostructural with zinc in zinc finger peptides and proteins, regardless of the number of cysteine residues involved. However, there is a strong, although quantitative variable thermodynamic preference for zinc over cobalt in tetrahedral environments provided by zinc finger. This effect is due to the ligand field stabilization effect, modulated by entropic contributions (Lachenmann et al., 2004). The substitution of nickel into zinc finger can also be achieved, but it results in distortions of the binding geometry and alterations of the peptide fold. A distorted tetrahedral coordination was found for cysteine2-Histidine2 zinc fingers (Posewitz & Wilcox, 1995), while a nearly square planar arrangement of donors was demonstrated in a cysteine4 environment (Bal et al., 2003). Also the efficacy of cadmium binding is related to the number of coordinated thiolates (Krizek et al., 1993). Zinc seems to be preferred in the cysteine2-histidine2 environment, the affinities of zinc and cadmium may be comparable for cysteine3-histidine, and cadmium is strongly preferred in cysteine4 zinc finger peptides (Kopera et al., 2004). This preference is due to the high enthalpy of the cadmium-S bond. The cadmium ion fits into tetrahedral environments with little strain, however, it is significantly larger than zinc and cobalt ions, which results in local distortions of zinc finger geometries (Buchko et al., 2000). Finally lead, element with high affinities to thiolates, can replace zinc in zinc finger domains and disrupt their fold. Arsenic too is known to have a high affinity to -SH groups and it is demonstrated by arsenic-mediated repair inhibition not related with a direct mechanism of one or more specific repair proteins, but rather with changes in gene expression and/or signal transduction (Asmuß et al., 2000). Finally redox regulation has been *in vitro* and *in vivo* demonstrated in several DNA-binding zinc finger

Different mechanism of action have been suggested for arsenic carcinogenicity including the induction of oxidative stress, induction of genetic damage, altered DNA methylation patterns, enhanced cell proliferation, inhibition of the tumor suppressor protein p53, DNA repair alteration and recently biomethylation (Aposhian & Aposhian 2006). A possible molecular mechanism for arsenic toxicity may lie in its ability to react with thiols, for example, in zinc binding structures of transcription factors, cell cycle control and DNA repair proteins (Kitchin & Wallace, 2008). Nucleotide excision repair (NER) in particular is strongly inhibited by arsenic. NER is capable of removing a wide variety of bulky, DNA helix distorting lesions, caused, e.g., by UV-irradiation or environmental mutagens. Arsenic is known to enhance the persistence of bulky DNA lesions and consequently the mutagenicity induced by UV and benzo[a]pyrene (Hartmann & Speit 1996; Gebel, 2001). Since bulky lesion formation is the possible responsible for their carcinogenicity, genetic integrity depends largely on NER efficiency. Many studies have shown that inorganic arsenic inhibits repair of bulky DNA adducts induced by UV-irradiation (Hartwig et al., 1997; Danaee et al., 2004) or benzo[a]pyrene in cultured cells and laboratory animals (Schwerdtle et al., 2003; Shen et al., 2008); additionally arsenite has been shown to downregulate expression of some NER genes in cultured human cells (Hamadeh et al., 2005). In

Interactions by Carcinogenic Metal Compounds with DNA Repair Processes 33

demonstrated to be highly time-dependent. Nollen et al. (2009) investigated the gene expression, total protein level and localization of proteins during NER and comparing inorganic arsenite and MMA(III). Arsenite and MMA(III) strongly decreased expression and protein level of the main initiator of global genome NER, i.e. Xeroderma pigmentosum complementation group C (XPC). This led to diminished association of XPC to sites of local UVC damage, resulting in decreased recruitment of further NER proteins. These data demonstrate that in human skin fibroblasts arsenite and MMA(III) more interacts with XPC expression, resulting in decreased XPC protein level and diminished assembly of the NER. The observed stronger impact on XPC by MMA(III) may explain the more potent NER inhibition by MMA(III) as compared to arsenite (Schwerdtle et al., 2003; Shen et al., 2008). Finally, these data provide further evidence that in the case of DNA repair inhibition the biomethylation of arsenic increases inorganic arsenic induced genotoxicity and probably contributes to its carcinogenicity. With respect to DNA repair inhibition, several studies point to an interaction of arsenic with various DNA repair pathways, which may in turn decrease genomic integrity. The effect of arsenic on the extent of poly(ADP-ribosyl)ation has been investigated previously in two studies with controversial conclusion. Yager & Wiencke (1997) observed a decreased amount of poly(ADP-ribose) in human T-cell lymphomaderived at arsenite concentrations above 5 µM. In contrast, an increase of poly(ADPribosyl)ation reaction was reported at higher concentrations in CHO-K1 cells (Lynn et al., 1998). Hartwig et al., 2003b investigated the effects of arsenite on poly(ADP-ribosyl)ation stimulated by H2O2 in intact cells by applying an anti-poly(ADP-ribose) monoclonal antibody. The experiments demonstrated a clear reduction of poly(ADP-ribosyl)ation level just evident at the extremely low and non-cytotoxic concentration of 10nM arsenite and reaching about 40% of residual activity at 0.5 µM arsenite. There was an increase in induced DNA single strand break formation by arsenite in agreement with the assumed role of poly(ADP-ribosyl)ation in DNA strand break repair (Hartwig et al., 2003b). Also the effect of the arsenicals on the activity of the isolated formamidopyrimidine glycosylase (Fpg) was examined. Fpg is a glycosylase initiating base excision repair in Escherichia coli: it recognizes and removes a lot of DNA base modification including 7,8-dihydro-8 oxoguanine (8-oxoguanine). Even though Fpg is a bacterial repair protein, the recent discovery of human homologues suggests its relevance for mammalian cells too (Hazra et al., 2003). After 30 min preincubation MMA(III) and DMA(III) inhibited Fpg activity in dosedependent shape, yielding 48 and 15% of the Fpg activity at 1 mM, respectively. In contrast, arsenite and the pentavalent metabolites did not show any effects on Fpg activity up to 10mM (Schwerdtle et al., 2003b). Finally, we describe the effects of arsenic compounds on the zinc finger structure of XPA. Different arsenicals promote the release of zinc from a peptide consisting of 37 amino acids representing the zinc finger domain of the human XPA protein (XPAzf). All trivalent arsenic compounds induced zinc release from XPAzf, starting at low micromolar concentrations, with MMA(III) and DMA(III) more active than arsenite. In contrast, MMA(V) and DMA(V) showed no or only slight effects up to 10mM (Schwerdtle et al., 2003b). Moreover there are some evidence about the influence of arsenic on BER, the predominant repair pathway for DNA lesions caused by reactive oxygen species (ROS) (Liu et al., 2001). Some studies have shown that low doses of arsenic can also cause an hormetic response in DNA polymerase β (Pol β), as well as telomerase activity (Zhang et al., 2003; Snow et al. 2005). DNA polymerase β is not only responsible for the incorporation of nucleotides in BER, but also excises the 5′-deoxyribose-5-phosphate (dRP) moiety prior to completion of repair (Wilson, 1998). Sykora et al. (2008) investigated the regulation of DNA

humans, arsenic exposure via drinking water was correlated with a dose relationship dependent to decreased expression of some NER genes and reduced repair of lesions in lymphocytes (Andrew et al., 2006). Human lymphoblastoid cells were pre-exposed to arsenite (As(III)) alone and in combination with UV, the pre-treatment with As(III) specifically inhibited the repair of UV-induced pyrimidine dimer-related DNA damage and leads to enhanced UV mutagenesis. Hartwig et al (1997) investigated the effects of arsenite in removal of benzo[a]pyrene-induced DNA damage. When damaged DNA is replicated prior to repair, these adducts can lead to mutations and cancer. This study was carried out in A549 human lung cancer cells; in absence of arsenite, about 45% of benzo[a]pyrene diolepoxide–DNA adducts were repaired within 6–8 h, in presence of arsenite, there was a significant increase of adduct formation. Additionally, the repair capacity towards the stable lesions was decreased in a concentration-dependent shape reaching about 25% of the control at 75 µM, a still slightly cytotoxic effect for this cell line (Schwerdtle et al., 2003b). Similar results have been obtained *in vivo*. Thus, in rats the frequency of benzo[a]pyrene-induced DNA adducts quantified by 32P post-labeling was drastically reduced in the presence of arsenite (Tran et al., 2002). Interesting was the evidence in the human study, arsenic exposure was associated with decreased expression of excision repair cross-complement 1 (ERCC1) in isolated human lymphocytes at the mRNA and protein levels. The mRNA levels of ERCC1 expression were positively associated with water arsenic concentration and nail arsenic concentration and significantly correlated with the amount of OGG1, a base pair excision repair gene (Mo et al., 2009). More detailed studies have been undertaken to assess the potential effects of the trivalent and pentavalent methylated metabolites on DNA repair processes. In humans and many other mammals, inorganic arsenic is converted into trivalent and pentavalent methylated metabolites, monomethylarsonous (MMA(III)) and dimethylarsinous (DMA(III)) acid, monomethylarsonic (MMA(V)) and dimethylarsinic (DMA(V)) acid. Biomethylation has long been thought to be a sort of detoxification process, yet nowadays it is reasonable to conclude that some adverse health effects seen in humans chronically exposed to inorganic arsenic are in fact caused by these metabolites. When considering MMA(III) and DMA(III) been demonstrated in some investigations as toxic, or even more toxic, compared to inorganic arsenic with an increase in benzo[a]pyrene diolepoxide–DNA adducts formation and repair inhibition for MMA(III), at much lower concentrations than arsenite (Schwerdtle et al., 2003). Repair inhibition was also observed at 5 µM DMA(III), but no effect on adduct generation was evident. Nevertheless, the cytotoxicity of the trivalent metabolites was also higher as compared to arsenite (Hartwig et al., 2003). Moreover, significant but less repair inhibition was mediated by 250 and 500µM of DMA(V) or MMA(V). Altogether, the results demonstrate that arsenite as well as the methylated metabolites interfere with cellular repair systems; the strongest effects with respect to inhibitory concentration were found for the trivalent metabolites (Schwerdtle et al., 2003b). Shen et al. (2009) investigated the difference manifested by DMA(III) compared to other trivalent arsenic species on the formation of benzo[a]pyrene diolepoxide–DNA adducts. At concentrations comparable to those used in the study by Schwerdtle et al. (2003) they found that each of the three trivalent arsenic species were able to enhance the formation of benzo[a]pyrene diolepoxide–DNA adducts with the potency in a decreasing order of MMA(III) > DMA(III) > As(III), well related with their cytotoxicity. Similar to As(III), DMA(III) the modulation of reduced glutathione (GSH) or total glutathione Stransferase (GST) activity could not account for its enhanced effect on DNA adduct formation. Additionally, similar effects elicited by the trivalent arsenic species were

humans, arsenic exposure via drinking water was correlated with a dose relationship dependent to decreased expression of some NER genes and reduced repair of lesions in lymphocytes (Andrew et al., 2006). Human lymphoblastoid cells were pre-exposed to arsenite (As(III)) alone and in combination with UV, the pre-treatment with As(III) specifically inhibited the repair of UV-induced pyrimidine dimer-related DNA damage and leads to enhanced UV mutagenesis. Hartwig et al (1997) investigated the effects of arsenite in removal of benzo[a]pyrene-induced DNA damage. When damaged DNA is replicated prior to repair, these adducts can lead to mutations and cancer. This study was carried out in A549 human lung cancer cells; in absence of arsenite, about 45% of benzo[a]pyrene diolepoxide–DNA adducts were repaired within 6–8 h, in presence of arsenite, there was a significant increase of adduct formation. Additionally, the repair capacity towards the stable lesions was decreased in a concentration-dependent shape reaching about 25% of the control at 75 µM, a still slightly cytotoxic effect for this cell line (Schwerdtle et al., 2003b). Similar results have been obtained *in vivo*. Thus, in rats the frequency of benzo[a]pyrene-induced DNA adducts quantified by 32P post-labeling was drastically reduced in the presence of arsenite (Tran et al., 2002). Interesting was the evidence in the human study, arsenic exposure was associated with decreased expression of excision repair cross-complement 1 (ERCC1) in isolated human lymphocytes at the mRNA and protein levels. The mRNA levels of ERCC1 expression were positively associated with water arsenic concentration and nail arsenic concentration and significantly correlated with the amount of OGG1, a base pair excision repair gene (Mo et al., 2009). More detailed studies have been undertaken to assess the potential effects of the trivalent and pentavalent methylated metabolites on DNA repair processes. In humans and many other mammals, inorganic arsenic is converted into trivalent and pentavalent methylated metabolites, monomethylarsonous (MMA(III)) and dimethylarsinous (DMA(III)) acid, monomethylarsonic (MMA(V)) and dimethylarsinic (DMA(V)) acid. Biomethylation has long been thought to be a sort of detoxification process, yet nowadays it is reasonable to conclude that some adverse health effects seen in humans chronically exposed to inorganic arsenic are in fact caused by these metabolites. When considering MMA(III) and DMA(III) been demonstrated in some investigations as toxic, or even more toxic, compared to inorganic arsenic with an increase in benzo[a]pyrene diolepoxide–DNA adducts formation and repair inhibition for MMA(III), at much lower concentrations than arsenite (Schwerdtle et al., 2003). Repair inhibition was also observed at 5 µM DMA(III), but no effect on adduct generation was evident. Nevertheless, the cytotoxicity of the trivalent metabolites was also higher as compared to arsenite (Hartwig et al., 2003). Moreover, significant but less repair inhibition was mediated by 250 and 500µM of DMA(V) or MMA(V). Altogether, the results demonstrate that arsenite as well as the methylated metabolites interfere with cellular repair systems; the strongest effects with respect to inhibitory concentration were found for the trivalent metabolites (Schwerdtle et al., 2003b). Shen et al. (2009) investigated the difference manifested by DMA(III) compared to other trivalent arsenic species on the formation of benzo[a]pyrene diolepoxide–DNA adducts. At concentrations comparable to those used in the study by Schwerdtle et al. (2003) they found that each of the three trivalent arsenic species were able to enhance the formation of benzo[a]pyrene diolepoxide–DNA adducts with the potency in a decreasing order of MMA(III) > DMA(III) > As(III), well related with their cytotoxicity. Similar to As(III), DMA(III) the modulation of reduced glutathione (GSH) or total glutathione Stransferase (GST) activity could not account for its enhanced effect on DNA adduct formation. Additionally, similar effects elicited by the trivalent arsenic species were demonstrated to be highly time-dependent. Nollen et al. (2009) investigated the gene expression, total protein level and localization of proteins during NER and comparing inorganic arsenite and MMA(III). Arsenite and MMA(III) strongly decreased expression and protein level of the main initiator of global genome NER, i.e. Xeroderma pigmentosum complementation group C (XPC). This led to diminished association of XPC to sites of local UVC damage, resulting in decreased recruitment of further NER proteins. These data demonstrate that in human skin fibroblasts arsenite and MMA(III) more interacts with XPC expression, resulting in decreased XPC protein level and diminished assembly of the NER. The observed stronger impact on XPC by MMA(III) may explain the more potent NER inhibition by MMA(III) as compared to arsenite (Schwerdtle et al., 2003; Shen et al., 2008). Finally, these data provide further evidence that in the case of DNA repair inhibition the biomethylation of arsenic increases inorganic arsenic induced genotoxicity and probably contributes to its carcinogenicity. With respect to DNA repair inhibition, several studies point to an interaction of arsenic with various DNA repair pathways, which may in turn decrease genomic integrity. The effect of arsenic on the extent of poly(ADP-ribosyl)ation has been investigated previously in two studies with controversial conclusion. Yager & Wiencke (1997) observed a decreased amount of poly(ADP-ribose) in human T-cell lymphomaderived at arsenite concentrations above 5 µM. In contrast, an increase of poly(ADPribosyl)ation reaction was reported at higher concentrations in CHO-K1 cells (Lynn et al., 1998). Hartwig et al., 2003b investigated the effects of arsenite on poly(ADP-ribosyl)ation stimulated by H2O2 in intact cells by applying an anti-poly(ADP-ribose) monoclonal antibody. The experiments demonstrated a clear reduction of poly(ADP-ribosyl)ation level just evident at the extremely low and non-cytotoxic concentration of 10nM arsenite and reaching about 40% of residual activity at 0.5 µM arsenite. There was an increase in induced DNA single strand break formation by arsenite in agreement with the assumed role of poly(ADP-ribosyl)ation in DNA strand break repair (Hartwig et al., 2003b). Also the effect of the arsenicals on the activity of the isolated formamidopyrimidine glycosylase (Fpg) was examined. Fpg is a glycosylase initiating base excision repair in Escherichia coli: it recognizes and removes a lot of DNA base modification including 7,8-dihydro-8 oxoguanine (8-oxoguanine). Even though Fpg is a bacterial repair protein, the recent discovery of human homologues suggests its relevance for mammalian cells too (Hazra et al., 2003). After 30 min preincubation MMA(III) and DMA(III) inhibited Fpg activity in dosedependent shape, yielding 48 and 15% of the Fpg activity at 1 mM, respectively. In contrast, arsenite and the pentavalent metabolites did not show any effects on Fpg activity up to 10mM (Schwerdtle et al., 2003b). Finally, we describe the effects of arsenic compounds on the zinc finger structure of XPA. Different arsenicals promote the release of zinc from a peptide consisting of 37 amino acids representing the zinc finger domain of the human XPA protein (XPAzf). All trivalent arsenic compounds induced zinc release from XPAzf, starting at low micromolar concentrations, with MMA(III) and DMA(III) more active than arsenite. In contrast, MMA(V) and DMA(V) showed no or only slight effects up to 10mM (Schwerdtle et al., 2003b). Moreover there are some evidence about the influence of arsenic on BER, the predominant repair pathway for DNA lesions caused by reactive oxygen species (ROS) (Liu et al., 2001). Some studies have shown that low doses of arsenic can also cause an hormetic response in DNA polymerase β (Pol β), as well as telomerase activity (Zhang et al., 2003; Snow et al. 2005). DNA polymerase β is not only responsible for the incorporation of nucleotides in BER, but also excises the 5′-deoxyribose-5-phosphate (dRP) moiety prior to completion of repair (Wilson, 1998). Sykora et al. (2008) investigated the regulation of DNA

Interactions by Carcinogenic Metal Compounds with DNA Repair Processes 35

on pH of medium and ions concentration. The DNA of rat primary hepatocytes was treated by incubation with 2-acetylaminofluorene, a known DNA damaging agent, and co-incubated with beryllium metal extracts (Strupp, 2011b). They observed that, the DNA repair synthesis were reduced by co-incubation with beryllium metal extract. However, it should be noted that this effect was observed only when the concurrent DNA damage was massive (>80% cells in repair), while no effects were observed in cells with lower DNA damage. These findings deserve however further investigations about their relevance in

Several reports suggested that cadmium genotoxicity is not direct but rather mediated by reactive oxygen free radicals and resulting oxidative stress. In spite of being a weak genotoxic chemical, cadmium exhibits remarkable potential to inhibit DNA damage repair, and this has been identified as a major mechanism for its carcinogenicity (Giaginis et al., 2006). Cadmium is comutagenic and increases the mutagenicity of UV radiaton, alkylation and oxidation in mammalian cells. These effects may be explained by cadmium inhibition on several types of DNA repair: base excision repair, nucleotide excision repair, mismatch repair and the elimination of the premutagenic DNA precursor 8-oxodGTP. Regarding base excision repair, low concentrations of cadmium which did not generate oxidative damage as such, inhibited the repair of oxidative DNA damage in mammalian cells (Dally & Hartwig 1997; Fatur et al. 2003). Exposure of human cells to sub-lethal concentrations of cadmium leads to a time and concentration dependent decrease in hOGG1 activity, i.e. of the main DNA glycosylase activity responsible for the initiation of the base excision repair of 8-oxoguanine, an abundant and mutagenic form of oxidized guanine. The study of Bravard et al. (2010) confirms that part of the inhibitory effect of low dose cadmium on the cellular 8-oxoguanine DNA glycosylase activity can be attributed to an already described reduced hOGG1 transcription (Youn et al., 2005). This moderate inhibitory effect of cadmium on hOGG1 mRNA levels cannot explain the dramatic decrease observed in the levels and activity of hOGG1 protein. Indeed, inhibition of the ectopically expressed hOGG1-GFP in cells exposed to the metal confirmed the post-transcriptional effect of cadmium on hOGG1 protein and activity levels. A different response of the second enzyme in the cellular BER pathway has been described. Bravard et al (2010) found that in vivo treatment of human cells with cadmium has no effect on the APE1 activity, suggesting that in their experimental conditions most cadmium is complexed within the cells and therefore the intracellular concentrations of free cadmium do not reach the levels required for the inhibition of APE1. These results, taken together with the indirect inhibition of hOGG1 by oxidation, support the hypothesis that the effects on the BER pathway are in the consequence of the cellular redox imbalance rather than the direct interaction with proteins. Candelas et al. (2010) showed that cadmium inhibits the repair of uracile (U) in DNA, resulting both from mis-incorporation and cytosine (C) deamination. These lesions, as those on AP sites, are common in any cell, and must constantly be repaired to avoid mutagenic events. The necessity to continuously repair these lesions is underscored by the high levels of expression of UNG2 and APE1 (Cappelli et al., 2001). This genotoxic consequence of cadmium exposure might participate in the deregulation of physiological cellular processes by altering the pattern of gene expression on the one hand (U), and increasing the mutation rate on the other hand (on

vivo.

**5. Cadmium** 

polymerase β (Pol β) and AP endonuclease (APE1), in response to low but physiologically relevant doses of arsenic. Lung fibroblasts and keratinocytes were exposed to As(III), and mRNA, protein levels and BER activity were assessed. Both Pol β and APE1 mRNA exhibited significant dose-dependent down regulation at doses of As(III) above 1 μM. However, at lower doses Pol β mRNA and protein levels, and consequently, BER activity were significantly increased. In contrast, APE1 protein levels were only marginally increased by low doses of As(III) and there was no correlation between APE1 and overall BER activity. Enzyme supplementation of nuclear extracts confirmed that Pol β was rate limiting. These changes in BER are related to the overall protective against sunlight UVinduced toxicity at low doses of As(III) while at high doses there is a synergistic toxicity action. The results provide evidence that changes in BER due to low doses of arsenic could contribute to a non-linear, threshold dose response for arsenic carcinogenesis. The primary function of APE1 in BER is to act as an endonuclease responsible for the excision of apurinic/apyrimidinic (AP) sites. However, APE1 is also a redox factor responsible for signal transduction in response to oxidative stress (Hsieh et al., 2001). Arsenic has the potential to affect both the endonuclease and the functions of APE1, through its increase in ROS levels and inhibition of DNA repair (Hamadeh et al., 2002).

Fig. 2. Schematic outline of DNA repair inhibition by arsenite and its methylated metabolites(modified from Hartwig et al., 2003).
