Abstract

MUTAGENESIS

REVIEW ARTICLE

Several pathways of hydrogen peroxide action that damage the E. coli genome

Nasser Ribeiro Asad^I; Lidia Maria Buarque Oliveira Asad^I; Carlos Eduardo Bonacossa de Almeida^III; Israel Felzenszwalb^I; Januário Bispo Cabral-Neto^II; Alvaro Costa Leitão^II

^IUniversidade do Estado do Rio de Janeiro, Departamento de Biofísica e Biometria, Instituto de Biologia Roberto Alcantara Gomes, Rio de Janeiro, RJ, Brazil

^IIUniversidade Federal do Rio de Janeiro, Instituto de Biofísica Carlos Chagas Filho, Rio de Janeiro, RJ, Brazil

^IIIComissão Nacional de Energia Nuclear, Divisão de Monitoração Individual, Instituto de Radioproteção e Dosimetria, Rio de Janeiro, RJ, Brazil

^{Correspondence} Correspondence to Dr. Alvaro C. Leitão Universidade Federal do Rio de Janeiro Instituto de Biofísica Carlos Chagas Filho, Centro de Ciências da Saúde, Bloco G Ilha do Fundão, 21949-900 Rio de Janeiro, RJ, Brasil E-mail: acleitao@biof.ufrj.br

ABSTRACT

Hydrogen peroxide is an important reactive oxygen species (ROS) that arises either during the aerobic respiration process or as a by-product of water radiolysis after exposure to ionizing radiation. The reaction of hydrogen peroxide with transition metals imposes on cells an oxidative stress condition that can result in damage to cell components such as proteins, lipids and principally to DNA, leading to mutagenesis and cell death. Escherichia coli cells are able to deal with these adverse events via DNA repair mechanisms, which enable them to recover their genome integrity. These include base excision repair (BER), nucleotide excision repair (NER) and recombinational repair. Other important defense mechanisms present in Escherichia coli are OxyR and SosRS anti-oxidant inducible pathways, which are elicited by cells to avoid the introduction of oxidative lesions by hydrogen peroxide. This review summarizes the phenomena of lethal synergism between UV irradiation (254 nm) and H₂O₂, the cross-adaptive response between different classes of genotoxic agents and hydrogen peroxide, and the role of copper ions in the lethal response to H₂O₂ under low-iron conditions.

Key words: hydrogen peroxide, cross-adaptive response, lethal synergism, copper and iron.

General Aspects

The appearance of aerobic forms of life was an important step in the evolutionary process, since oxygen consumption leads to the production of ten-fold more energy from glucose than does anaerobic metabolism (Meneghini, 1987). However, this process imposes constraints on cell viability, because of the generation of reactive oxygen species during respiration.

The consecutive univalent reduction of molecular oxygen to water produces three active intermediates: superoxide anion (O₂^-•), hydrogen peroxide (H₂O₂) and hydroxyl radical (OH^•). These intermediates, collectively referred to as reactive oxygen species (ROS) are potent oxidants of lipids, proteins, and nucleic acids (Halliwell and Gutteridge, 1984; Mello-Filho and Meneghini, 1985; Meneghini, 1988). Among the oxidative DNA lesions, one of the major classes of DNA damage leads to modification in purine and pyrimidine bases, together with oligonucleotide strand breaks, DNA-protein cross-links and abasic sites. Increasing evidence suggests that the cumulative damage caused by ROS contributes to numerous degenerative diseases associated with aging, such as atherosclerosis, rheumatoid arthritis and cancer (Ames et al., 1993; Halliwell and Gutteridge, 1999).

Living organisms have developed specific mechanisms to prevent the production and effects of ROS. The reduction of O₂ by cytochrome oxidase without yielding ROS, the superoxide dismutase catalysis of O₂^-• into H₂O₂ through a dismutation reaction, the decomposition of H₂O₂ by catalase and peroxidases, and the scavenging of ROS by some vitamins comprise part of the set of cellular antioxidant defenses (Halliwell and Gutteridge, 1999). Synthesis of the enzymes that catalyze these reactions is a part of the adaptive response triggered by the stress posed by ROS.

Despite these cellular defenses, many lesions can be produced in cellular targets, mainly in the DNA molecule, leading to mutagenesis and cell death. However, cells are able to deal with these adverse events via DNA repair mechanisms, which enable them to recover their integrity. These include base excision repair (BER), nucleotide excision repair (NER) and recombinational repair (Miles and Sancar, 1989).

Several studies indicate that the killing of E. coli cells exposed to H₂O₂ is mainly due to damage to DNA (Imlay and Linn, 1986; Imlay and Linn, 1988; Hagensee and Moses, 1989; Asad and Leitão, 1991) and a wide variety of DNA lesions are formed (Halliwell and Gutteridge, 1999) as a by-product of the H₂O₂ reaction. Some of these lesions are miscoding acting as an important source of mutagenesis in aerobically growing cells.

Participation of OH^• as the main damaging agent has been suggested by studies using scavengers of OH^•. Repine et al. (1981) and Brakely et al. (1990) have demonstrated that dimethyl sulfoxide partially inhibits DNA base damage by H₂O₂, and Brandi et al. (1989) have noted that E. coli bacteria are partially protected against the lethal effects of H₂O₂ by pretreatment with ethanol, dimethyl sulfoxide, or thiourea.

The induction of DNA damage by H₂O₂ in E. coli as well as in mammalian cells can be either impaired or enhanced by the presence of transition metal ion chelators (Mello-Filho and Meneghini, 1985: Asad and Leitão, 1991). In addition, H₂O₂ can cause membrane lesions through lipid peroxidation and, by promoting alterations in several amino acids can lead to the inactivation of enzymes (Farr and Kogona 1991).

Recent reviews on the mechanisms of oxidative DNA damage and repair have focussed on prevention and repair (Cooke et al., 2003), biochemical features (Cadet et al., 2003), substrate specificities for glycosylases (Dizdaroglu, 2003), biological consequences (Wallace, 2002: Bjelland and Seeberg, 2003) and the role of iron (Kruszewski, 2003).

In this review, we discuss several aspects of DNA damage and cellular inactivation induced by hydrogen peroxide; the importance of transition metals such as iron and copper in this context; DNA repair pathways involved in the cellular response to H₂O₂, and related antioxidant cell defenses.

H₂O₂ and transition metals

H₂O₂ reacts with O₂^-• resulting in OH^• production through the so-called Haber-Weis cycle. However, this reaction depends on the presence of transition metals such as Cu⁺ and/or Fe²⁺, which work as reducing agents, according to the reactions below (Kehrer, 2000).

H₂O₂, per se, is considered a weak oxidant agent. Nevertheless, it easily crosses the cellular membrane and reacts with transition metals, generating OH^•. Evidence for the importance of transition metals comes from studies with chelators such as dipyridyl, phenanthroline and desferioxamine (Fe chelators) and neocuproine (a Cu chelator) (Brandi et al., 1989; Brakely et al., 1990; Hallywell and Gutteridge, 1999). Pretreatment with iron chelators protects both prokaryotic and eukaryotic cells against the lethal effects of H₂O₂ (Mello-Filho and Meneghini, 1985; Imlay and Linn, 1988; Asad and Leitão, 1991). Additionally, cultures of Staphylococcus aureus are more sensitive to H₂O₂ when iron is added to the culture media (Repine et al., 1981) and in E. coli cultures the same is observed (Touati et al., 1995).

In E. coli the metabolism of iron is involved in the cellular antioxidant response. A transcription factor denoted Fur, the global repressor of ferric ion uptake, regulates about 30 genes implicated in iron uptake from the environment (Braun, 1997; Braun et al., 1998). Most Fur-regulated genes are derepressed in low iron concentrations and repressed when a high concentration of iron is present (Hantke, 1981; Hantke, 2002). The finding that fur mutants are sensitive to H₂O₂ and that they suffer an increase in oxidative DNA damage leading to mutations under aerobic conditions supports the hypothesis that Fur has a role in the defense against oxidative stress (Touati et al., 1995). Furthermore, the regulators of E. coli responses to oxidative stress, OxyR and SoxRS, activate the expression of Fur and suggest that control of iron metabolism in E. coli is an integral part of the antioxidant response (Zheng et al., 1999)

DNA Repair Pathways and Antioxidant Defense Systems

Base excision repair (BER)

Lesions produced by H₂O₂ are typically repaired by base excision repair (BER) mechanisms. This kind of DNA repair is initiated by DNA glycosylases, enzymes that recognize the modified bases and act by cleaving the glycosylic bond, thereby removing the damaged base from the sugar phosphate backbone and, as a result, producing an apurinic/apyrimidinic (AP) site (Sancar and Sancar, 1988).

Some DNA glycosylases also display a class I AP lyase activity that incises the phosphodiester linkage on the 3' side of the AP lesion and generates a 5'-phosphate group and a 3'-terminus that needs removal by a class II AP endonuclease/3'-diesterase prior to repair synthesis and ligation (Piersen et al., 2000). The cleavage of AP sites can also be catalyzed by class II AP endonucleases which incise the 5' side of the AP site, leaving a 3'-OH terminus and a 5'-abasic residue that is removed by a deoxyribophosphodiesterase (dRPase) (Mol et al., 2000).

In E. coli two enzymes are representative of the class I AP lyases, the Fpg/MutM DNA glycosylase, product of the fpg/mutM gene, which recognizes primarily oxidized purines, and the DNA endonuclease III, product of the nth gene, which recognizes primarily oxidized pirimidines. Interestingly, the E. coli repair-deficient fpg/mutM, as well as the nth mutants, are not more sensitive to killing by H₂O₂ than wild-type cells (Cunningham and Weiss, 1985; Boiteux and Huisman, 1989). Subsequently, endonuclease VIII and IX were purified from endonuclease III-deficient E. coli cells (Wallace et al., 1988; Melamede et al., 1994). Endonuclease VIII was found to exhibit a thymine glycol DNA glycosylase activity as well as an AP lyase activity. Endonuclease IX recognizes urea residues and b-ureidoisobutyric acid in DNA; however, DNA containing thymine glycol or dihydrothymine is not a substrate for this enzyme (Friedberg et al., 1995). Later, the gene for endonuclease VIII (nei) was isolated (Jiang et al., 1997; Saito et al., 1997). Endonuclease VIII is present in extracts of E. coli at 5 to 10% of the level of endonuclease III and is responsible for the repair of 10% of the thymine glycol in E. coli (Wallace et al., 1988; Wallace, 1988).

The nth nei double mutants are hypersensitive to H₂O₂ (Jiang et al., 1997). Moreover, endonuclease VIII can recognize 8-oxo-7,8-dihydroxyguanine (8-oxoG) lesions, a kind of lesion also recognized by Fpg protein (Blaisdell et al., 1999). The multiplicity of DNA glycosylases that recognize and attack sites of oxidative damage in DNA confirms the importance of this form of base damage.

Class II AP endonucleases of E. coli are mainly represented by the exonuclease III, the xthA gene product, and endonuclease IV, the nfo gene product.

The role of exonuclease III in repairing oxidative damage was highlighted by the demonstration that xth mutants are extremely sensitive to H₂O₂ (Demple et al., 1983). On the other hand, nfo mutants are not sensitive to H₂O₂ (Cunningham et al., 1986), and deletion of nfo increases the killing of xth mutants to H₂O₂, indicating that many of the repair activities of exonuclease III and endonuclease IV overlap.

Endonuclease IV normally represents about 10% of the total endonuclease activity; superoxide-generating agents induce a 10- to 20-fold increase in the level of this enzyme through the soxRS response (Chan and Weiss, 1987).

Nucleotide excision repair (NER) in response to H₂O₂

In E. coli, a complex of proteins encoded by the uvrA, uvrB and uvrC genes is required for lesion recognition and the dual incisions. This complex eliminates DNA lesions that cause significant distortions in the phosphodiester backbone of the molecule (Friedberg et al., 1995; Hanawalt, 2001).

This kind of repair is of fundamental importance for the correction of UV (254 nm) lesions, mainly cyclobutane pyrimidine dimers. It appears not to be related to the repair of H₂O₂ lesions, since the uvrABC mutants are not sensitive to H₂O₂ (Imlay and Linn, 1987).

The finding that the triple mutant uvrA nfo xthA cannot be constructed, despite the fact that the double mutants are viable (Saporito et al., 1989), has raised the question of the connection between BER and NER DNA repair pathways. Additionally, the UvrABC complex is able to remove AP sites generated by oxidative lesions in the DNA molecule (Lin and Sancar, 1989). So, these results suggest that both repair systems correct the lesions produced by oxidant agents produced during cellular respiration.

These findings suggest that some intermediate products of base excision repair may be substrates for the UvrABC complex. This hypothesis was confirmed by Kow et al. (1990), who demonstrated the role of this complex in the repair of thymine glycols in the DNA of replicative form of phage FX174. The survival of phage containing thymine glycols is lower in the nth uvrA double mutant than in the nth single mutant.

SOS response

The best-studied transcriptional response to DNA damage is the SOS response (Friedberg et al., 1995; Walker, 1996). Single-stranded DNA produced by several DNA-damaging agents and repair mechanisms can be bound by RecA protein, resulting in conversion of this protein to its activated form. Once activated, RecA interacts with LexA protein, the repressor of the SOS genes (Wagner et al., 1999). This interaction triggers the autocatalytic cleavage of LexA and consequent destruction of its ability to function as a repressor, which results in the derepression of SOS genes (Mustard and Little, 2000; Fernandez De Henestrosa et al., 2000). By using DNA microarray techniques Courcelle et al. (2001) have shown that in E. coli the expression of 43 genes is controlled by LexA.

The expression of LexA-controlled genes allows the increased phenotypic expression of mutagenesis (umuDC genes), nucleotide excision repair (uvrA and uvrB genes), genetic recombination (recA, recN, recQ and recD genes), cellular filamentation (sulA and sulB genes), and survival and mutagenesis of irradiated phages (W-reactivation and W-mutagenesis) (Friedberg et al., 1995).

A low concentration of H₂O₂ (1-3 mM) results in SOS gene induction in wild-type cells (Imlay and Linn, 1987; Goerlich et al., 1989). However, H₂O₂ can induce some SOS responses without SOS induction. An example is the cell filamentation induced by H₂O₂ in sulA and recA mutants and the mutagenesis that occurs at the same level in wild-type and umuC mutant cells treated with H₂O₂ (Imlay and Linn, 1987). On the other hand, the induction of SOS by H₂O₂ is an important event, since recA and recBC mutant cells are very sensitive to H₂O₂ treatment probably due to the lack of recombinational repair necessary for the repair of H₂O₂-induced lesions (Imlay and Linn, 1987). Additionally, Konola et al., (2000) have shown that the ruvA mutants are 10- to 15-fold more sensitive to H₂O₂ (1-3 mM) than the wild-type cells. Together with RuvB, the RuvA protein generates the RuvAB complex, which stimulates strand migration in the Holliday junctions (West, 1996).

Other inducible events: anti-oxidant and cross-adaptive responses triggered by H₂O₂

Most genes encoding DNA repair enzymes that act on oxidative damage appear to be expressed constitutively in actively growing cells. This is presumably because oxidative DNA damage is continuously produced by ROS, which are normal by-products of aerobic metabolism (Demple and Harrison, 1994; Henle and Linn, 1997). However, in order to deal with elevated levels of peroxide in the environment, cells have evolved changes in metabolism that help to protect DNA from ROS.

Most inducible genes that respond to oxidative damage prevent, rather than repair DNA damage. However, a notable exception is endonuclease IV, the nfo gene product, an AP endonuclease that repairs 3' phosphate residues to 3'OH groups that can prime DNA synthesis (Chaudhry et al., 1999; Izumi et al., 2000).

Two key protective responses have been described in E. coli - one controlled by soxRS genes and the other by oxyR (Tsaneva and Weiss, 1990; Storz and Imlay, 1999; Gonzalez-Flecha and Demple, 2000).

The SoxRS

Low concentrations of superoxide-generating compounds such as paraquat and menadione render the cells resistant to higher doses of these agents (Geenberg and Demple, 1989) in a manner dependent on the integrity of the soxRS locus.

The soxRS regulatory system acts in two steps, with SoxR serving both as a sensor and as an activator protein. When activated by the univalent oxidation of the 2Fe-2S clusters of the protein through a not yet explained mechanism (Storz and Imlay, 1999) SoxR induces transcription of soxS, a positive regulator that stimulates transcription of more than 16 other superoxide responsive genes (Wu and Weiss, 1992; Hidalgo et al., 1995). Although this system responds to oxidative stress when cells are exposed to superoxide radical-generating agents, it is not induced by H₂O₂ (Chan and Weiss, 1987; Tsaneva and Weiss, 1990; Hidalgo et al., 1997). However, it was demonstrated that in some conditions H₂O₂ as well as singlet molecular oxygen could activate the SoxRS regulon in vivo (Manchado et al., 2000; Agnez-Lima et al., 2001).

The products of the induced soxRS regulon include: Mn-superoxide dismutase (sodA), DNA repair endonuclease IV (nfo), glucose-6-phosphate dehydrogenase (zwf), aconitase (acnA), stable fumarase (fumC), ferredoxin reductase (fpr), toxin and antibiotic efflux pumps (acrAB), an antisense RNA for the ompF porin mRNA (micF) and an iron-binding repressor of iron transport (fur) (Amabile-Cuevas and Demple, 1991;Gaudu and Weiss, 1996; Gaudu et al., 1997; Pomposiello and Demple, 2001).

Regulation of the OxyRS response to oxidative damage

Bacterial cells possess an adaptive response to oxidizing agents, which means that exposure to low levels of H₂O₂ allows bacterial cells to survive further toxic doses of H₂O₂ (Demple and Halbrook, 1983; Demple, 1991). The expression of nine proteins induced by H₂O₂ treatment is under the control of the oxyR gene (Christman et al., 1985). Several proteins whose expression is regulated by oxyR have been identified, including catalase and an alkyl hydroperoxide reductase (Morgan et al., 1986; Storz et al., 1990).

The level of OxyR protein does not change with H₂O₂ treatment, indicating that it is activated post-translationally (Storz et al., 1990). H₂O₂ activates the transcriptional activity of OxyR by oxidizing two of its cysteine residues (Zheng et al., 1998; Aslung et al., 1999; Storz and Toledano, 1999). When activated, OxyR activates transcription of genes that include katG (catalase hydroperoxidase I), ahpCF (alkylhidroperoxide-NADPH oxido-reductase), grxA (glutaredoxin), gorA (glutathione reductase), dps (a protein that protects DNA from peroxide damage) (Altuvia et al., 1994; Martinez and Kolter, 1997) and fur (an iron-binding repressor of iron transport). Under oxidative-stress conditions with the influence of OxyR and SoxRS, the number of Fur molecules per cell increases from 5,000 to about 10,000 (Zheng et al., 1999).

In addition, OxyR activates the synthesis of oxyS (Altuvia et al., 1997; Zhang et al., 1997), which encodes an untranslated mRNA that appears to regulate as many as 20 additional genes, possibly by an antisense mechanism (Altuvia et al., 1997; Argaman and Altuvia, 2000). The oxyS gene product is independent of OxyR in limiting the endogenous production of H₂O₂ in E. coli (Gonzalez-Flexa and Demple 1999).

Through random transcriptional fusions some new genes, the expression of which require OxyR, have been detected. These include henF (coproporphyrinogen III oxidase), which participates in the synthesis of photoheme IX, which is required for activity of both HPI (katG) and HPII (Mukhopadhyay and Schellhorn, 1997). DNA microarray techniques have detected several other new OxyR-activated genes including the henH heme biosynthetic gene; the six-gene suf operon, which may participate in Fe-S cluster assembly of repair; and four genes of unknown function (Zheng et al., 2001).

Cross-Adaptive Responses

Cross-adaptive response occurs when cells exposed to doses of a sub-lethal agent develop resistance against challenging doses of another lethal agent. For instance, it is well known that E. coli cells exposed to low doses of H₂O₂ develop resistance against heat shock, ethanol (Jenkins et al., 1988), ultraviolet A (UVA) (Tyrrell, 1985), formaldehyde (Nunoshiba et al., 1991), menadione and cumene hydroperoxide (Christman et al., 1985). In contrast, prior exposure of E. coli to low doses of H₂O₂ has shown little or no effect on the resistance of these cells to UVC or alkylating agents (Demple and Halbrook, 1983). This section is meant to provide a general view into the cross-adaptive response induced by H₂O₂ against effects of UV (254 nm), N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) and cumene hydroperoxide.

Cross-adaptive response between H₂O₂ and UV (254 nm)

Demple and Halbrook (1983) have shown that prior exposure of E. coli to low doses of H₂O₂, in the micromolar range, engenders little or no effect on the resistance of these cells to UV. However, pretreatment with 2.5 mM H₂O₂ protected wild-type cells against UV-irradiation. This protection is independent of the SOS response, since it is also observed in a lexA mutant, and is dependent on DNA excision repair, since this protection is not observed in an uvrA mutant (Asad et al., 1994). The possibility that DNA repair activities other than those known to be involved in SOS response could be induced during H₂O₂ pretreatment. was verified by examining the ability of H₂O₂-pretreated cells to repair incoming damaged DNA. In this way, experiments similar to those described by Weigle (1953), but involving treatment of the cells with 2.5 mM H₂O₂ instead of UV, were carried out with UV-irradiated phages. An enhanced survival of UV-irradiated phages in the wild-type, H₂O₂-pretreated cells has been shown. The same results were obtained with the lexA1 mutant, indicating that the response is not dependent on the SOS system; nevertheless RecA and UvrA proteins are involved, since there is no UV-damaged phage reactivation in recA and uvrA mutant cells pretreated with 2.5 mM H₂O₂. The oxyD3 mutant, in which the adaptive response to H₂O₂ does not occur (Imlay and Linn, 1987), presented similar results to those observed in wild-type and the lexA1 mutant, indicating that the oxyR regulon does not participate in the UV-damaged phage repair which takes place in H₂O₂-treated cells.

The induction of some proteins after UV irradiation in lexA (Def) mutants has been observed (Lesca et al., 1991), and these proteins may act in mutagenesis and in the DNA repair of UV-irradiated bacteriophage lambda (Calsou et al., 1987). In fact, Petit et al. (1993) have characterized an E. coli gene, dinY, whose induction does not require the cleavage of LexA repressor, although it may be considered a member of the SOS regulon (Petit et al., 1993; Friedberg et al., 1995). The protective effect induced by H₂O₂ against UV is shown to be independent of the induction of dinY gene or other genes under the same control as dinY (Asad et al., 2000). Besides, this cross-protection response is not induced when the SOS regulon is constitutively expressed. The most consistent view of these observations seems to be that the induction of this response may be related to the induction of genes that are not under the control of LexA, and are inhibited by the expression of SOS genes (Asad et al., 2000).

Cross-adaptive response between H₂O₂ and MNNG

The Ada and Ogt proteins are involved in the DNA repair of some lesions caused by alkylating agents, as O⁶-methylguanine and O⁴-methylthymine. Moreover, Ada induces the expression of different genes such as ada, alkA and aidB in cells treated with alkylating agents, leading to the development of adaptive responses induced by pretreatment with sublethal concentrations of alkylating agents (Friedberg et al., 1995).

A cross-adaptive response against lethal effects caused by alkylating agents does not occur when cells of E. coli are exposed to low doses of H₂O₂ (micromolar) (Demple and Halbrook, 1983), but on the other hand, pretreatment with 5 mM H₂O₂ protects the wild-type strain, as well as ada, ogt, ada-ogt, alkA and aidB mutants against the lethal effect of MNNG (Assad et al., 1997). Since H₂O₂ is able to oxidize thiols (Halliwell and Gutteridge, 1999), which are necessary to convert MNNG into the mutagenic methylnitrosamine (Sedgwick and Robins, 1980), similar experiments were performed with N-nitroso-N-methylurea and N-nitroso-N-ethylurea, which do not require activation by thiols, and the results were similar to those obtained with MNNG (Assad et al., 1997). This protection is accompanied by a reduction in the mutation frequency in the wild-type cells and in the ogt mutant, but not in the ada mutant. So, Ada protein is able to decrease the mutagenic lesions induced by MNNG in H₂O₂-pretreated cells. Horsfall et al. (1990) have shown that the DNA context influences both the distribution and reparability of alkylation damage. This observation may provide information on the cross-adaptive response between H₂O₂ and MNNG, since the DNA alkylation pattern induced by MNNG can be altered when the DNA has already been oxidized.

Cross-adaptive response between H₂O₂ and cumene hydroperoxide

Ahp protein expression is under the control of OxyR protein. Ahp plays an important role in protecting bacterial cells against alkyl hydroperoxides, such as cumene hydroperoxide (Storz et al., 1989), and the active enzyme requires the presence of two subunits with molecular weights of 22 kDa (AhpC) and 57 kDa (AhpF).

It has been shown that 2.5 mM H₂O₂ pre-treatment protects wild-type cells against cumene hydroperoxide killing (Asad et al., 1998). Unexpectedly, this protection is observed in the oxyD 3 mutants, meaning that it is an OxyR-independent phenomenon. The protection is mediated by Ahp protein and there is no requirement for novel protein synthesis (Asad et al., 1998). Electrophoretic profile studies of proteins from wild-type cells treated with H₂O₂ (2.5 mM) for 20 min have shown that a 22 kDa protein (possibly Ahp protein) observed in untreated wild-type cells is faintly visible in H₂O₂-treated cells. In the latter, instead of the 22 kDa band a 24 kDa band appears. This modification was not observed when wild type was treated with lower concentrations of H₂O₂ (micromolar).

It is known that H₂O₂ causes modification in several E. coli proteins (Farr and Kogona, 1991). So, these results suggest that millimolar concentrations of H₂O₂ might induce an alteration in the electrophoretic profile of the smaller subunit of AhpC (22 kDa). It is important to bear in mind that inhibition of protein synthesis does not prevent the appearance of the 24-kDa band promoted by H₂O₂. These results are interpreted to mean that AhpC 22-kDa subunit undergoes conversion to different oxidized forms when challenged with higher doses of H₂O₂, resulting in a small alteration of its molecular weight. Since Ahp may prevent oxidative damage and repair lesions caused by active oxygen species (Farr and Kogona, 1991), the AhpC 24 kDa form can be interfering with the repair of lesions caused by cumene hydroperoxide (Asad et al., 1998).

DNA damage and repair in E coli challenged by H₂O₂ in the presence of iron chelators

The protection conferred by iron scavengers against the effect of H₂O₂ was first demonstrated by Imlay and Linn (1988). In 1991, Asad and Leitão showed that prior treatment with iron chelators such as o-phenanthroline, dipyridyl and desferrioxamine protected the cells against the H₂O₂ lethal effect, but the number of induced DNA-single strand breaks (DNA-SSB) was similar even in the presence of iron chelators. More significantly, the breaks observed after treatment with metal chelators and H₂O₂ are repaired 60 min after H₂O₂ elimination in xthA but not polA mutant cells (Asad and Leitão, 1991).

Since then, in a series of experiments spanning a full decade, DNA damage and the genetic repair mechanisms of the H₂O₂-induced DNA lesions in a low-iron condition, which means cells previously treated with metal-ion chelators, have been examined. Asad et al. (1995) demonstrated a striking hypersensitivity of the fpg and uvrA mutant strains to H₂O₂ under low-iron conditions, suggesting that the processing of DNA lesions induced in this situation may occur in a different way from that found under physiological iron conditions.

Other investigations have demonstrated common pathways in the cell response to H₂O₂ challenge in a low-iron condition and in physiological iron conditions. SOS and the OxyR pathways are the ones involved in the response of E. coli cells subjected to both challenges (Asad et al., 1997).

Evidence for the participation of endonuclease IV (the nfo gene product) in the repair of DNA lesions generated by H₂O₂ under low-iron conditions came from studies of the processing of DNA strand breaks in different E. coli strains (Galhardo et al., 2000). Survival experiments with xthA, nfo, and xthA nfo mutant strains in which the cultures were treated with 5 mM H₂O₂ confirm data previously described in the literature, so that the xthA mutant was sensitive to H₂O₂ (Demple et al., 1983) while the nfo strain was not (Cunningham et al., 1986), and the xthA nfo double mutant was even more sensitive than the xthA mutant. Experiments with cultures pretreated with dipyridyl showed a distinct pattern of dependence on AP endonucleases for bacterial survival. Neither the xthA nor the nfo single mutants were significantly inactivated by H₂O₂ under low-iron conditions. However, the xthA nfo double mutant was highly sensitive to this treatment, indicating that both exonuclease III and endonuclease IV may act in the repair of the oxidative lesions generated under such conditions. Analysis of the sedimentation profile in alkaline sucrose gradients also demonstrated that both xthA and nfo mutants, but not the xthA nfo double mutant, could carry out complete repair of DNA-SSB generated by H₂O₂ under low-iron conditions (Galhardo et al., 2000). Thus, these findings support the idea that both exonuclease III and endonuclease IV act in the repair of DNA damage induced by H₂O₂ in iron-depleted E. coli, and then strongly suggest that the lesions caused by H₂O₂ in the presence of dipyridyl are qualitatively different from those found in the absence of this iron chelator. It is interesting to note that despite the severe repair defect of the xthA nfo strain, significant levels of repair were observed under both physiological and low-iron conditions. The exact nature of this repair mechanism remains to be elucidated.

Further studies indicated that the formation of substrates for exonuclease III and endonuclease IV is mediated by the Fpg DNA glycosylase, since the fpg mutation increases cell survival and repair of DNA strand breaks in a null AP endonuclease background (xthA nfo double mutant). Recently, Speck et al. (2002) showed that the same phenomenon is observed for nitric oxide treatment, that is, in a null AP endonuclease background, fpg and ung mutations are able to increase cell survival.

The role of copper ions in the lethality induced by H₂O₂ in low-iron conditions

Most of the work on the effects of H₂O₂ in living organisms has reported that DNA damage induced by H₂O₂ can be explained basically by the generation of hydroxyl radicals through the iron-mediated Fenton reaction. Some authors have attributed the same role to copper ions, also present in biological systems (Sagripanti and Kraemer, 1989; Aruoma et al., 1991; Dizdaroglu et al., 1991). The suggested mechanism for this effect takes into account the formation of a DNA-Cu²⁺ complex in which this metal would be reduced and then the DNA-Cu⁺ would react with H₂O₂ to produce oxidative damage, via Fenton-type reactions (Aruoma et al., 1991; Byrnes et al., 1992; Lloyd et al., 1997; Lloyd and Phillips, 1999). Although copper ions seemed not to participate in the genotoxicity of H₂O₂ in E. coli under physiological iron conditions (Asad and Leitão, 1991), these ions have been shown to take part in the genotoxicity of H₂O₂ under conditions of low iron availability. In fact, neocuproine (copper ion chelator) can inhibit cell inactivation and DNA strand breakage caused by H₂O₂ in the presence of iron chelators (Almeida et al., 2000). This phenomenon can only be detected in high concentrations of H₂O₂ (15 mM), suggesting that these ions only interact with H₂O₂ in the intracellular environment in the absence of iron and under severe oxidative stress.

Copper-induced DNA damage has been studied in several systems. It has been demonstrated that such damage is targeted preferentially to adjacent polyguanosines (Aruoma et al., 1991). Analysis of the oxidative base lesions generated by copper-mediated Fenton reactions showed that purine residues are the preferred targets of the DNA-damaging species (Sagripanti and Kraemer 1989; Dizdaroglu et al., 1991; Lloyd and Phillips, 1999], with 8-oxoG being the most abundant lesion formed. Frelon et al. (2003) have shown that only 8-oxoG is formed upon incubation of DNA with Cu(II) ions and H₂O₂ and suggested that in vitro the Fenton reaction triggered by copper ions generates singlet oxygen as the predominant reactive species, with hydroxyl radical being produced predominantly when the Fenton reaction is triggered by iron ions. In fact, 8-oxoG is the only lesion excised from H₂O₂/Cu²⁺-treated DNA at detectable levels by the yOgg1 glycosylase from Saccharomyces cerevisiae (Karahalil et al., 1998). The finding of a remarkable increase in mutagenesis in uvrA fpg-strain cultures pretreated with dipyridyl, and its inhibition by neocuproine (Almeida et al., 2000), strongly suggest that 8-oxoG is formed in large amounts by such treatment. The data suggesting a significant role for Fpg and UvrA proteins in the repair of DNA lesions correlate well with the hypothesis of copper participation, since it was demonstrated that these proteins are important in the repair of lesions induced by singlet oxygen, another well-known guanine-damaging agent. Taken together these results may indicate the production of singlet oxygen as the predominant ROS when E. coli cultures are challenged with H₂O₂ under low-iron conditions.

On the other hand some interesting results were obtained from mutagenesis assays performed by our group (manuscript in preparation) regarding the nature of lesions produced by H₂O₂ in cells previously treated with dipyridyl. Using an assay based on lac^- reversion through a single base change in mutated lacZ codon (Cupples and Miller, 1989), we found that concentrations of H₂O₂ above 15 mM induce almost exclusively A:T® T:A transversions. However, in cells previously treated with dipyridyl a massive and significant presence of G:C® A:T transitions is detected, a clearly different profile of induced mutations. Considering that this transversion is not reported to be induced by 8-oxoG the authors have suggested that H₂O₂ under low-iron conditions may generate the lesion 5-hydroxy-2'-deoxycytidine, a highly mutagenic product of cytosine oxidation that also constitutes a substrate for Fpg repair protein (Hatahet et al., 1994; Feig et al., 1994).

A qualitative difference exists between the DNA lesions generated by H₂O₂ plus copper (in the presence of iron chelators) and iron (physiologic iron condition), which can be deduced by the different DNA repair requirements for cell survival in these situations. The reason for this difference is not known, and there exists an as yet unsolved question about the participation of hydroxyl radical in the damage to DNA caused by copper plus H₂O₂ (Sagripanti and Kraemer, 1989; Yamamoto and Kawanishi, 1989). Free-radical scavengers partially inhibit E. coli inactivation by H₂O₂ (Imlay et al., 1988), and we observed that thiourea protects the cells against the lethality produced by H₂O₂ both in the presence of dipyridyl (our unpublished results) and in physiological iron conditions (Asad and Leitão, 1991).

Synergistic lethal interactions

In the last three decades, many studies have been published regarding synergistic lethal effects between H₂O₂ and different physical and chemical agents. Some of them will be described in this section.

Lethal synergism between UV (254 nm) and H₂O₂

Synergistic lethal interaction between UV (254 nm) and X-ray irradiation, a well known system of generating H₂O₂ and DNA strand breaks was described three decades ago in E. coli (Haynes, 1966; Martignoni and Smith, 1973). However, synergistic lethal interaction was not observed between UV (254 nm) and H₂O₂ (Hartman and Eisenstark, 1980). On the other hand, synergistic killing of E. coli and S. thyphimurium by near-UV (300-400 nm) radiation and H₂O₂ has been described (Hartman and Eisenstark, 1978; Hartman and Eisenstark, 1980; Kramer and Ames, 1987).

An analysis of the UV and X-ray effects (Martignoni and Smith, 1973), the results of repair of H₂O₂ lesions in several mutant strains (Ananthaswany and Eisenstark, 1977), and our own results for UV and reductone (an H₂O₂-generating keto-aldehyde) (Leitão et al., 1981a; Leitão et al., 1981b) suggested that a synergistic lethal interaction between UV and H₂O₂ (254 nm) should be observed. Indeed, Leitão and Carvalho (1988) observed that prior UV (254 nm) irradiation strongly increased the sensitivity to H₂O₂ of wild-type E. coli cells, and this synergistic lethal interaction was also observed to a reduced extent in a polA mutant, suggesting that UV lesions are potentiated by the additional damage produced by H₂O₂, a result similar to those observed by Martignoni and Smith (1973) for UV and X-rays.

The detection of DNA double-strand breaks by DNA sedimentation on neutral sucrose gradients (Leitão and Carvalho, 1988) clearly indicates a mechanism responsible for the synergistic lethal interaction observed. Although H₂O₂ can produce DNA strand breaks with low efficiency (Demple and Linn, 1982), the repair of UV lesions by the action of the uvr gene products generated single-stranded DNA regions and it was supposed that, in the presence of H₂O₂, DNA double-strand breaks might arise in these regions, produced by the action of exonuclease III (the xthA gene product). This hypothesis was confirmed by the absence of synergistic lethal interaction between UV (254 nm) and H₂O₂ in xthA as well as in uvrA mutant strains (Leitão and Carvalho, 1988). It seems that the same kind of mechanism may be operating in the synergisms observed with near-UV and H₂O₂, UV and reductone and UV and X-rays.

Lethal synergism between phenanthrolines and H₂O₂

As described earlier, prior treatment with iron chelators such as desferrioxamine, dipyridyl or o-phenanthroline (1,10-phenanthroline) protects E. coli cells against the lethal effects of H₂O₂. However, Asad et al. (1994) detected a strong lethal interaction when xthA mutant cells were treated simultaneously with H₂O₂ and o-phenanthroline. In the same way Almeida et al. (1999) have also detected a lethal synergistic interaction between neocuproine (2,9-dimethyl 1,10-phenanthroline) and H₂O₂. In both cases the phenomenon of synergism was accompanied by an increase in the number of DNA strand breaks.

In the case of the synergism described by Asad et al. (1994), it was argued that the formation and dissociation equilibrium of Fe²⁺-Phe complex, which follows equations IV, V and VI (Lee et al., 1948) shown below, would explain the observed phenomenon.

Since the equilibrium of the bis and mono complexes of Fe²⁺-Phe are rapidly established and since these bis and mono complexes react quickly with H₂O₂ (Burgers and Prince, 1965), there would be H₂O₂-mono and H₂O₂-bis complex formation and these complexes may be extremely lethal to the cells, therefore justifying the lethal interaction observed. In this case, the ability of o-phenanthroline to penetrate into the DNA duplex, acting as a shuttle for the Fe²⁺ ions would increase the efficiency of OH generation close to DNA (Furtado et al., 1997). Recently Furtado (2002) showed that this interaction also occurs in vitro. By using plasmid DNA, he showed that the high number of breaks, as measured by the transformation of supercoiled to relaxed form, is obtained when iron and o-phenanthroline are added to the reaction medium immediately before H₂O₂.

As mentioned above, Almeida et al. (1999) showed that prior incubation of fpg, uvrA and lexA mutant strains with neocuproine led to an increased sensitivity to H₂O₂. The chemistry of this synergistic lethal interaction was suggested to occur via the iron-mediated Fenton reaction, which would be responsible for displacement of copper ions from the complex Neo₂Cu⁺, as Florence et al., (1985) have reported to occur in vitro. In fact, neocuproine, as well as other phenanthroline derivatives, have the ability to penetrate into DNA, guiding metal ions to this site, which can contribute to radical formation in the vicinity of this molecule (Furtado et al., 1997). Indeed, it is possible that neocuproine can guide copper ions to the DNA molecule, thereby promoting the occurrence of radical generation at this site as H₂O₂ reacts with these ions.

Conclusions

The most remarkable outcome from the genetic studies on H₂O₂-mediated genotoxicity is the striking sensitivity of xthA mutant E. coli cells to H₂O₂. In fact, this observation provides some insights into the nature of H₂O₂-induced lesions. Nevertheless, it is important to realize that the oxidative stress produced by H₂O₂ results in the induction of a diverse set of physiological responses, which include some paradoxical effects. In this context, the most unexpected phenomenon investigated so far is the bimodal pattern of inactivation of E. coli by H₂O₂ described by Imlay and Linn (1986). In the same way, it is now evident that although pre-treatment of E. coli xthA cultures with iron chelators confers protection against the lethal effects of H₂O₂, DNA lesions can still be formed under these conditions. We have to keep in mind that in contrast with what is observed in the repair of damage induced by H₂O₂ under physiological iron conditions, after treatment with H₂O₂ in low-iron conditions: (i) the lesions observed may be repaired in xthA mutant cells (ii) both endonuclease IV and exonuclease III as well as Fpg and UvrA proteins participate in the repair of these lesions. Such a difference in the repair suggests a qualitative difference in the formation of DNA lesions that is independent of the presence of iron. Iron chelators partially inhibit E. coli inactivation by H₂O₂. However, the precise mechanism for the genotoxic effect of H₂O₂ under low-iron conditions remains to be elucidated. It is also curious to see that some metal-ion chelators such as o-phenanthroline and neocuproine, which should generally inhibit the lethal effect of H₂O₂ can sometimes enhance cell inactivation. Moreover, the data available so far suggest that H₂O₂ (2.5 mM) can induce protection against UV light, MNNG and cumene hydroperoxide independently of the adaptive response, indicating that H₂O₂ can produce cross-protection responses through various mechanisms that might not involve the induction of de novo protein synthesis.

Acknowledgements

We thank C.F.M. Menck, D.P. Carvalho and M. Sorenson for critical reading of this manuscript. Research supported by CNPq, CAPES, FUJB, FAPERJ and PRONEX.

EditorAssociado: Carlos F.M. Menck

Received: February 13, 2004;

Accepted: February 19, 2004.

Publication Dates

History