Mutation Research
Fundamental and Molecular Mechanisms of Mutagenesis
684:24-34 (2010).
Chloroacetaldehyde-induced mutagenesis in Escherichia coli: the role of AlkB protein in repair of 3,N4-ethenocytosine and 3,N4-α-hydroxyethanocytosine
Agnieszka M. Maciejewskaa, Karol P. Ruszela, Jadwiga Nieminuszczya, Joanna Lewickaa, Beata Sokołowskab, Elżbieta Grzesiuka and Jarosław T. Kuśmiereka,*
aInstitute of Biochemistry and Biophysics, Polish Academy of Sciences, 02-106 Warsaw, 5A Pawińskiego Str, Poland; bMedical Research Center, Polish Academy of Sciences, 02-106 Warsaw, 5 Pawińskiego Str, Poland
* To whom correspondence should be addressed. Tel.: +48 22592 3338; Fax: +48 22592 2190; E-mail: .
Keywords: E. coli AlkB, AlkA and Mug proteins; repair of etheno adducts; chloroacetaldehyde-induced mutagenesis; LacZ→Lac+ reversion
Abbreviations: CAA, chloroacetaldehyde; ε, etheno, see Fig. 1 for structures and abbreviations of etheno adducts; MF, mutation frequency; MMS, methyl methanesulphonate.
.
Abstract
Etheno () adducts are formed in reaction of DNA bases with various environmental carcinogens and endogenously created products of lipid peroxidation. Chloroacetaldehyde (CAA), a metabolite of carcinogen vinyl chloride, is routinely used to generate -adducts. We studied the role of AlkB, along with AlkA and Mug proteins, all engaged in repair of -adducts, in CAA-induced mutagenesis. The test system used involved pIF102 and pIF104 plasmids bearing the lactose operon of CC102 or CC104 origin (C.G. Cupples, J.H. Miller. Proc. Natl. Acad. Sci. U.S.A. 86 (1989) 5345-5349) which allowed to monitor Lac+ revertants, the latter arose by GCAT or GCTA substitutions, respectively, as a result of modification of guanine and cytosine. The plasmids were CAA-damaged in vitro and replicated in E. coli of various genetic backgrounds. To modify the levels of AlkA and AlkB proteins, mutagenesis was studied in E. coli cells induced or not in adaptive response. Formation of C proceeds viaa relatively stable intermediate, 3,N4--hydroxyethanocytosine (HEC), which allowed to compare repair of both adducts. The results indicate that all three genes, alkA, alkB and mug, are engaged in alleviation of CAA-induced mutagenesis. The frequency of mutation was higher in AlkA-, AlkB- and Mug-deficient strains in comparison to alkA+, alkB+,and mug+ controls. Considering the levels of CAA-induced Lac+ revertants in strains harboring the pIF plasmids and induced or not in adaptive response, we conclude that AlkB protein is engaged in the repair of C and HEC in vivo. Using the modified TTCTT 5-mers as substrates,we confirmed in vitro that AlkB protein repairs C and HEC although far less efficiently than the reference adduct 3-methylcytosine. The pH optimum for repair of HEC and εC is significantly different from that for 3-methylcytosine. We propose that the protonated form of adduct interact in active site of AlkB protein.
1. Introduction
The alkB gene was discovered over twenty years ago in Mutsuo Sekiguchi’s laboratory while isolating Escherichia coli mutants with increased sensitivity towards alkylating agents [1]. This gene is a part of the system of adaptive response to alkylating agents in E. coli (the Ada response), in which induced alkylation resistance results from increased expression of four genes, ada, alkB, alkA and aidB. The Ada, AlkA and AlkB proteins protect against alkylation by repair of alkylated bases in DNA using different mechanisms, whereas the function of the aidB gene product has not been conclusively established (recently reviewed in [2]). Although the phenotypic characteristics of alkB mutants as well as logical reasoning indicated that 3-methylcytosine (m3C) and 1-methyladenine (m1A) are the most plausible substrates for AlkB [3], the direct biochemical evidence emerged only recently. Crucial information in elucidation of the enzymatic function of AlkB was provided by bioinformatics analysis which established that AlkB is a member of the superfamily of 2-oxoglutarate- and iron-dependent dioxygenases [4]. Independent studies in Barbara Sedgwick’s and Erling Seeberg’s laboratories have revealed that purified AlkB removes methyl groups from m3C and m1A via oxidation to formaldehyde in the presence of 2-oxoglutarate, Fe(II) and O2, thus restoring undamaged bases in DNA [5,6]. Further studies revealed that analogous alkB genes are present in various organisms including humans, and that besides methylated ssDNA, also dsDNA and RNA can be substrates for AlkB and its homologues. However, each protein displays a different repair propensity for particular type of nucleic acid. In addition to m3C and m1A, the structurally similar lesions, 1-methylguanine and 3-methylthymine are also repaired, albeit less efficiently, by human and bacterial AlkB proteins. Moreover, it was found that AlkB protein removes from DNA not only methyl groups but also higher alkyl homologues(for latest reviews see: [2,7-9]).
The etheno (ε) base adducts (see Fig. 1 for structures) constitute another, structurally distinct from linear alkyl base adducts, group of AlkB substrates. They belong to exocyclic DNA base adducts formed by various bifunctional agents of exogenous and endogenous origin. They are generated by many industrial carcinogens: e.g., vinyl chloride metabolites, chloroethylene epoxide and chloroacetaldehyde (CAA).The oxidative stress driven lipid peroxidation is a powerful source of endogenous reactive agents, which react with DNA to form etheno and other types of adducts, and accumulating evidence suggests that these types of lesions play a role in the etiology of some human cancers [10].
It has been found that the toxic and mutagenic effect caused by the presence of εA and εC in site-specifically modified M13 viral DNA in E. colialkB mutant was alleviated in alkB+ cells. Also, in vitro experiments have shown that the etheno bridge of εA, situated in an oligodeoxynucleotide, is removed by AlkB protein via epoxidation of the double bond and is finally released as glyoxal [11]. Similarly, the in vitro repair of εA by bacterial AlkB and its human homologue hABH3 has been demonstrated by another group [12]. 1,N6-ethanoadenine, an analogue of εA, containing a saturated two-carbon bridge formed through the reaction of adenine in DNA with the antitumor agent 1,3-bis(2-chloroethyl-1-nitrosourea (BCNU), upon action of AlkB protein is transformed into a non-toxic and weakly mutagenic N6-2-oxoethyl derivative [13]. AlkB also influences CAA-induced mutation spectra and toxicity in the pSP189 sup F shuttle vector, which implies AlkB contribution to repair of ε-adducts and other exocyclic adducts generated by CAA [14]. Recently, it has been shown that hABH2, but not hABH3, is active in the direct reversal of εA, and that mABH2 is the principal dioxygenase for εA repair in mice [15].
CAA is highly toxic and weakly mutagenic agent. This CAA feature makes studying CAA mutagenesis in bacteria very difficult. Here, using a new method that allows to examine the mutagenic potency of CAA but eliminates its cytotoxic action on the cell, we studied the role of AlkB protein in CAA-induced mutagenesis and repair of ethenoadducts in DNA. The test system used involved pIF102 and pIF104 plasmids [16] containing the lactose operon of the CC102 or CC104 strain [17]. The plasmid DNA was modified with CAA in vitro, but DNA repair processes were studied in vivo in live bacterial cells. Transformation of CAA-pretreated plasmid DNA into E. coli cells allowed to study GCAT transitions (pIF102) and GCTA transversions (pIF104) that can arise by the modification of both components of the G·C pair, guanine and cytosine. The GCAT transition dominates among base substitutions caused by CAA in E. coli cells[18]. Similarly, GCAT transitions followed by GCTA transversions are the most frequent base substitutions in CAA-treated shuttle vector propagated in the cultured human fibroblast cells[19]. It is worth to mention that GCAT transitions are the most frequent mutations in MMS-treated alkB strains [20]. Since the reaction of cytosine moiety with CAA proceeds via a relatively stable intermediate, 3,N4--hydroxyethanocytosine (HEC, Fig. 1), we attempted to compare the mutagenic potential and repair of both cytosine adducts, HEC and a stable product of its dehydration, εC.
The use of wild type E. coli and mutants deficient in AlkB, AlkA and the mismatch uracil-DNA glycosylase Mug, all engaged in repair of ε-adducts, allowed to estimate the role of each enzyme in DNA repair.We also present results indicating that purified AlkB protein repairs CAA-modified cytosines in oligodeoxynucleotides.
2. Materials and Methods
2.1. Bacterial strains
The E. coli K12 strains used in this study are listed in Table 1. The strains made for mutagenesis tests were constructed by P1-mediated transduction according to [21]. The AM1 strain was a recA+ DH5α derivative constructed by transduction of srl::Tn10, recA+ from EC1651 and selected for tetracycline resistance and loss of UV sensitivity. The obtained strain (DH5α recA+, srl::Tn10) was a recipient of srl+ gene from the AM100 strain (with selection for sorbitol as the only source of carbon). AM4 was also constructed in two steps. First, Tn10 zee3129, his was transferred from EC2445 to AM1 (with selection for tetracycline resistance followed by screening for histidine requirement). Then, to the resultant AM1 Tn10 zee3129, his strain, the alkA1 point mutation was co-transduced with the his+ gene from the MS23 strain. Transductants were selected on glucose minimal plates, and subsequently screened for loss of tetracycline resistance and enhanced MMS sensitivity. Details of construction of all AM1 derivative strains are given in Table 1. The alkB117 and mug-1 were selected on LB plates [21] containing 50 μg/mL carbenicillin or 50 μg/mL kanamycin, respectively. At every step the transductants were checkedfor enhanced MMS and CAA sensitivity.
2.2. Media
Bacteria were cultured in rich LB [21] and SOC [25] media. Minimal medium (MM) was C-salts [26] supplemented with 0.4% lactose, or 0.5% glucose, or 0.5 % sorbitol and thiamine (10 g/ml) if necessary. LB and MM were solidified by the addition of 1.5% Difco agar.
2.3. CAA modification of pIF102 and pIF104 plasmids
Plasmids were isolated according to standard protocol [25] and then treated with CAA at 37°C for 30 min. The incubation mixture in the volume of 0.4 mL contained: 100 μg of plasmid DNA, 300 mM sodium cacodylate buffer pH 7.5, and 2.9 – 29 μL CAA (Merck, 45% aq. solution, 6.88 M, final concentration 50 – 500 mM, respectively). After modification DNA was precipitated twice with 2.5 volumes of ice-cold ethanol in the presence of 0.3 M sodium acetate buffer pH 5.2, washed with cold ethanol, redissolved in appropriate volume of TE buffer and then frozen at -20ºC. The CAA-modified plasmids prepared this way contained the majority of cytosine-CAA adduct in the form of HEC (plasmid pIF102-HEC and pIF104-HEC). To convert the HEC intermediate to εC, the plasmid solutions were brought to pH 6.5 by addition of 0.8 M sodium cacodylate buffer pH 6.5 to a final concentration of 20 mM and incubated at 37°C for 72 h (plasmid pIF102-εC and pIF104-εC) [27]. Mock treated plasmids were prepared in the same way except that CAA was absent from the incubation mixture. The mutagenicity assay experiments comparing HEC- and εC-containing plasmids were always performed in parallel using plasmids from the same batch of CAA modification.
2.4. Mutagenicity assay: induction of adaptive response, preparation of electrocompetent cells, transformation by electroporation
Bacterial strains were cultured in two repeats (37ºC, anaerobic conditions, 250rpm). When the cultures reached OD600 = 0.4, to one of them MMS (Sigma) was added to a final concentration of 7.5 mM in order to induce Ada response. Subsequently both cultures were incubated for 15 min and then were used for preparation of electrocompetent cells according to standard procedure [25]. The competent cells (50 μL) and 0.5 μg of modified plasmid DNA were electroporated in a 2-mm ice-cold cuvette at 2.5 kV/cm using the default settings of an Eppendorf Electroporator 2510 (Eppendorf, Madison, WI). The recovery medium SOC was added to a 1mL final volume and then the transformation mixture was incubated at 37°C for 2 h under anaerobic conditions. Appropriate dilutions of the mixture were spread on LB plates containing chloramphenicol (30 μg/mL) to count the total number of transformed cells and on lactose MM plates for selectionof mutants. Plates were incubated at 37°C and counted after 24 and 72 h for transformation efficiency and for mutagenesis, respectively. Transformation efficiency was defined as the number of colony forming units (cfu) produced by 1 μg of plasmid DNA in a transformation reaction. Mutation frequency (MF) was calculated as the number of mutants per 104 transformed cells. The results of mutagenicity assays for CAA-treated plasmids pIF102 and pIF104 are visualized in Figs 2 and 3, respectively, and are given numerically in Supplementary Tables 1-4 (see Supplementary Materials).
2.5. Statistics
Data were presented as mean and standard deviation (SD). Analysis of variance (ANOVA) on ranks with Kruskal-Wallis test was applied. Differences between the groups were compared using the Mann-Whitney U test. The statistical evaluation was performed using a commercial statistical package (Statistica 5.1, StatSoft, Poland).
2.6. CAA and MMS modification of TTCTT oligodeoxynucleotide (T/C 5-mer)
Modification with CAA.To a solution of 220 nmol of the T/C 5-mer (Metabion, Martinsried, Germany) in 176 μL of 1.3 M triethylammonium bicarbonate (pH ~7.5) 16.4 μL of 6.88 M CAA (45% aq. solution, Merck) was added, and the mixture was incubated at 37°C for 3 h. The reaction mixture containing unreacted T/C, T/HEC and T/C 5-mers was purified from CAA and reaction buffer, and fractionated using preparative HPLC (see Section 2.10). The early eluting fractions containing T/C and T/HEC 5-mers were lyophilized, redissolved in TE buffer and used as HEC-substrate (85.7 % of T/HEC 5-mer) for the AlkB assay. Fractions containing a mixture of modified and unmodified T/C 5-mers, obtained in the other CAA reaction, were combined, lyophilized, redissolved and incubated (37°C, pH 6.5, 72 h) for conversion of HEC to εC (see Section 2.3). This material was used as εC-substrate (94.5 % of T/C 5-mer) for AlkB assay. The representative analytical HPLC profiles of HEC- and εC-substrates are shown in Fig. 4. Notice that the work-up procedure (lyophilization, storage at - 20°C, etc.) of HEC-substrate does not produce an appreciable amount of T/εC 5-mer. The identity of HEC- and εC-substrate was ascertained by HPLC analysis after enzymatic hydrolysis to deoxynucleosides according to[27] (not shown).
Modification with MMS.To a solution of 2.5 nmol of the T/C 5-mer in 52 μL of 0.77 M triethylammonium acetate pH 6.5, 2 μL of MMS (Sigma) was added, and the mixture was incubated at 37°C for 18 h. The mixture (72% of T/m3C 5-mer by HPLC analysis, not shown) was used as m3C-substrate for AlkB assay without any further purification.
2.7. AlkB dioxygenase purification
His-tagged E. coli AlkB protein was overexpressed in E. coli BL21(DE3) (Ivitrogen) carrying pET28balkB plasmid (a generous gift from the late prof. E. Seeberg laboratory (University of Oslo, Norway)) in 2 L of bacterial culture in LB medium. Expression was induced with 0.5 mM IPTG for 4 h. Bacterial pellets were sonicated in ice-cold binding buffer (20 mM sodium phosphate buffer pH 7.4, 300 mM NaCl, 10 mM imidazole, 10 mM -mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride) and centrifuged. The extracts were loaded onto a 1 ml Ni-NTA column (HiTrap Chelating HP, Amersham – Pharmacia Biotech), and eluted by imidazole gradient with final imidazole concentration of 500 mM. Fractions containing purified protein were dialyzed into 20mM sodium acetate buffer pH 5.0, containing 300 mM NaCl and 1 mM DTT; glycerol was subsequently added to final concentration of 50% and the protein was stored in -70°C.
2.9. AlkB assay
Our observations and also those of others [6] indicate that the presence or absence of ascorbate in the reaction mixture does not influence the repair activity of AlkB. Since ascorbate produces a massive peak in HPLC profile, we omitted this reagent in our assays.
HEC- and εC-substrate. The reaction mixtures in a volume of 20 μL contained: 35 mM sodium acetate buffer pH 3.8-5.0 (the most efficient repair determined after 30 min of reaction was at pH 4.6), 1 mM dithiotreithol, 0.2 mM Fe(NH4)2(SO4)2, 0.35 mM α-ketoglutarate, 500 pmoles of CAA-modified T/C 5-mer substrate and 50 pmoles of purified AlkB protein. Reactions were carried out at 37°C for various periods of time and then analyzed by HPLC. The representative separations are shown in Fig. 4. The pH-dependence and the time course of repair are shown in Fig. 5A and 5B, respectively.
m3C-substrate. The reaction mixtures in 20 μL volume contained: 50 mM Tris/HCl buffer pH 7.2-8.8 (the most efficient repair determined after 30 min of reaction was at pH 8.0), 1 mM dithiotreitol, 0.05 mM Fe(NH4)2(SO4)2, 0.5 mM α-ketoglutarate, 500 pmoles of m3C-substrate and 6.25 pmole of purified AlkB protein.Reactions were carried out and analyzed by HPLC as in the case of HEC- and εC-substrates. The pH-dependence and the time course of repair are shown in Fig. 5A and 5B, respectively.
2.10. HPLC chromatography
HPLC was performed using a Waters dual pump system with a tunable UV/visible light absorbance detector managed by Millenium 2010SS (version 2.15) controller. Analytical and preparative separations were performed, respectively, on a Waters Nova-Pak® C18, 60 Ǻ,4 μm, 4.6 × 250 mm, cartridge column, at a flow rate of 1 mL/min, and on HR C18, 60 Ǻ, 6 μm, 3.9 × 300mm, at a flow rate of 1.75 mL/min. In both cases a linear gradient of 20 mM triethylammonium acetate pH 6.5 × 30% aq. MeOH over 30 min was applied and detection was carried out at 270 nm.
3. Results
3.1. DNA reaction with CAA – the estimation of level of base modification.
The plasmid DNA was CAA-modified essentially under the same conditions as described by Kim et al.[14]. Therefore, we expect similar level of modification of plasmid DNA as in the cited work. The number of adducts per 100 nucleotides for 0.6 M CAA dosage and 1h reaction determined by LC-MS/MS method was as follows: 1,N2G – 0.5, C – 1.5 and A – 2.5. The authors did not measure the level of N2,3G. Our former studies indicate however, that in dsDNA N2,3G is formed with efficiency comparable to that of A [28]. Also in that work we have found that C is formed with highest efficiency among the etheno-adducted bases, whereas Kim et al.[14]reported lower efficiency of C than A. The discrepancy between these results comes probably from the fact that LC-MS/MS method did not include the hydrated form of C, namely HEC[14]. In our study CAA-modified DNA wasacid-depurinated, what caused simultaneous conversion of HEC to C, then enzymaticaly hydrolyzed. The resulted pyrimidine deoxynucleosides were analyzed using HPLC with UV detection, therefore the determined level of C represented the sum of initially formed HEC and product of its dehydration, C [28].
Under physiological conditions the half-life of HEC is ~1 day [29]. HEC is enough stable to detect it in enzymatic hydrolysates of CAA-treated oligonucleotides[27]. As it has been mentioned in Section 2.6 the work-up procedure (lyophilization, storage at - 20°C, etc.) of TT(HEC)TT does not produce an appreciable amount of TT(εC)TT 5-mer. The complete conversion of HEC to εC can be achieved by 72h incubation at pH 6.5 in 37º C[27]. Therefore, we used pIF102 and pIF104 plasmids freshly reacted with CAA (or stored frozen at -20ºC) which presumably contained the majority of cytosine-CAA adduct in the form of HEC (pIF102-HEC and pIF104-HEC) and the same plasmids with HEC dehydrated to C by prolonged incubation at pH 6.5 (pIF102-C and pIF104-C).
3.2. Transformation efficiency and calculation of mutation frequency