Mutagenic Spectrum of Butadiene-Derived N1-Deoxyinosine Adducts

Nov 19, 2002 - ... pollutant, being found in automobile exhaust and tobacco smoke. ...... C. J. (1996) Assessment of 1,3-butadiene mutagenicity in the...
0 downloads 0 Views 166KB Size
1572

Chem. Res. Toxicol. 2002, 15, 1572-1580

Mutagenic Spectrum of Butadiene-Derived N1-Deoxyinosine Adducts and N6,N6-Deoxyadenosine Intrastrand Cross-Links in Mammalian Cells Manorama Kanuri,† Lubomir V. Nechev,‡,§ Pamela J. Tamura,‡ Constance M. Harris,‡ Thomas M. Harris,‡ and R. Stephen Lloyd*,†,

|

Sealy Center for Molecular Science, University of Texas Medical Branch, Galveston, Texas 77555-1071, and Department of Chemistry, Vanderbilt University, Nashville, Tennessee 37235 Received August 2, 2002

Reactive metabolites of 1,3-butadiene, including 1,2-epoxy-3-butene (BDO), 1,2:3,4-diepoxybutane (BDO2), and 3,4-epoxy-1,2-butanediol (BDE), form both stable and unstable base adducts in DNA and have been implicated in producing genotoxic effects in rodents and human cells. N1 deoxyadenosine adducts are unstable and can undergo either hydrolytic deamination to yield N1 deoxyinosine adducts or Dimroth rearrangement to yield N6 adducts. The dominant point mutation observed at AT sites in both in vivo and in vitro mutagenesis studies using BD and its epoxides has been A f T transversions followed by A f G transitions. To understand which of the butadiene adducts are responsible for mutations at AT sites, the present study focuses on the N1 deoxyinosine adduct at C2 of BDO and N6,N6-deoxyadenosine intrastrand cross-links derived from BDO2. These lesions were incorporated site-specifically and stereospecifically into oligodeoxynucleotides which were engineered into mammalian shuttle vectors for replication bypass and mutational analyses in COS-7 cells. Replication of DNAs containing the R,R-BDO2 intrastrand cross-link between N6 positions of deoxyadenosine yielded a high frequency (59%) of single base substitutions at the 3′ adducted base, while 19% mutagenesis was detected using the S,S-diastereomer. Comparable studies using the R- and S-diastereomers of the N1 deoxyinosine adduct gave rise to ∼50 and 80% A f G transitions with overall mutagenic frequencies of 59 and 90%, respectively. Collectively, these data establish a molecular basis for A f G transitions that are observed following in vivo and in vitro exposures to BD and its epoxides, but fail to reveal the source of the A f T transversions that are the dominant point mutation.

Introduction Butadiene (BD,1 CH2dCHCHdCH2, Scheme 1) is a chemically reactive gas widely used in the manufacture of polymers such as styrene-butadiene and acrylonitrilebutadiene-styrene. It ranks among the top 20 highvolume chemicals produced in the United States (1) and is an environmental pollutant, being found in automobile exhaust and tobacco smoke. Butadiene is carcinogenic in mice, but less so in rats, and is classified as a human carcinogen in a recent report from the National Toxicology Program (2). The toxicology and epidemiology of butadiene have been previously reviewed (3, 4). The genetic effects of butadiene are attributed to its metabolic activation by cytochrome P450s (primarily 2A6 and 2E1) to the R- and * To whom the correspondence should be addressed: Phone: (409) 772- 2179. Fax: (409) 772-1790. E-mail: [email protected]. † University of Texas Medical Branch. ‡ Vanderbilt University. § Present address: Transgenomic, Inc., Boulder, CO 80301. | Dr. R. S. Lloyd holds the Mary Gibbs Jones Distinguished Chair in Environmental Toxicology from the Houston Endowment. 1 Abbreviations: BD, butadiene; BDO, butadiene monoepoxide (1,2epoxy-3-butene); BDO2, butadiene diepoxide (1,2:3,4-diepoxybutane); BDE, 3,4-epoxy-1,2-butanediol; FAB MS, fast atom bombardment mass spectrometry; MALDI-TOF, matrix-assisted laser desorption mass spectrometry.

Scheme 1. Reactive Metabolites of Butadiene

a BD, butadiene; BDO, butadiene monoepoxide, 1,2-epoxy-3butene; BDO2, butadiene diepoxide, 1,2,3.4-diepoxybutane; BDE, 3,4-epoxy-1,2-butanediol.

S-monoepoxides (BDO), which can undergo diverse reactions including further oxidation to give stereoisomers of butadiene diepoxide (BDO2). BDO2 can undergo hydrolysis to form the epoxide of butene-3,4-diol (BDE), which can also be formed by hydrolysis of BDO followed by further oxidation. BDO, BDO2, and BDE react with deoxynucleosides and DNA to form many different types of adducts (5-13). Examples include adducts at N1, N2, and N7 of deoxyguanosine, and N1 and N3 of deoxyadenosine and N3 adducts of deoxycytidine. The N1 adducts of deoxyadenosine can undergo Dimroth rear-

10.1021/tx025591g CCC: $22.00 © 2002 American Chemical Society Published on Web 11/19/2002

Mutagenesis of Butadiene-Induced Adducts

Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1573

Scheme 2. Formation of R- and S-Diastereomers of BDO-Deoxyinosine and R,R- and S,S-Diastereomers of N 6-N 6-BDO2 Cross-Links by Reaction of BDO and BDO2 with Deoxyadenosine

rangement to N 6 adducts or deamination to form the N1 deoxyinosine derivatives (Scheme 2). The N3 adducts of deoxycytidine deaminate to form deoxyuridine derivatives. To further complicate the problem, BDO can react at either the terminal (C1) or internal (C2) end of the epoxide and both types of adducts have been found. Without doubt, other adducts are also formed but have not yet been characterized. Thus far, the only adduct identified in samples from human subjects is the N1(2,3,4-trihydroxybutyl) derivative of adenine (14). BDO2 is formed in vivo in much smaller quantities than BDO but is much more genotoxic (15). The latter observation is generally attributed to the fact that it is a bis-electrophile and capable of cross-linking DNA to itself or to proteins. BDO2 is suspected of being the causative agent in the high incidence of lymphatic and hematopoietic cancers among chemical workers who are occupationally exposed to butadiene (16, 17). Large interspecies differences in carcinogenicity of butadiene have been observed and are attributed to differences in the ability to metabolize BDO to BDO2 (18, 19). No bis-DNA adducts of BDO2 have been detected in vivo; however, guanine N7-N7 cross-links have been isolated from treatment of salmon sperm DNA with BDO2 and inferred from reaction of BDO2 with oligodeoxynucleotides (20, 21). Concurrent with the identification and quantitation of DNA-bound BD metabolites, numerous studies have characterized the types of DNA damage induced by BD and its epoxides. In 1990, Goodrow et al. (22) examined NIH 3T3 cells transformed by DNA from lymphomas and lung and liver tumors induced in B6C3F1 mice by longterm exposure to BD. The most common mutation found was a G f C transversion in codon 13 of the K-ras protooncogene. Studies by Cochrane and Skopek (2325) of hprt mutations in TK6 human lymphoblastoid cells and splenic T cells exposed to BDO, BDO2, and BDE showed that BDO2 is approximately 100-fold more mutagenic than BDO and BDE was the least mutagenic of the three. Transitions and transversions at both GC and AT sites were detected, along with frame shift mutations. BDO2-treated cells showed large deletions, which had been seen also in Drosophila melanogaster exposed to BDO2 (26, 27). AT base pair mutations induced at the hprt locus in rodents and humans have been used as

biomarkers to determine whether there are genotoxic effects associated with exposures to butadiene (28, 29) and these data indicate a strong correlation between exposure and an increased frequency of lymphocytes with mutations in the hprt gene. Extensive in vivo mutagenesis studies in B6C3F1 lacI transgenic mice have been carried out by Recio (30-40). Mutations in the lacI gene resulting from exposure to BD were primarily point mutations (61% at GC and 20% at AT) (33). When Rat2 lacI transgenic fibroblasts were treated with BDO, the observed mutations were primarily base substitutions (60% at GC and 31% at AT) (31), whereas BDO2 in the same system induced micronuclei but few mutations (39). Failure to increase lacI mutation frequency was attributed to the inability of this lambda shuttle vector-based system to detect large deletions. When TK6 cells were exposed to BDO, most (78%) of the mutations observed in the hprt gene were single-base substitutions, with a significant level of A f T transversions (37). A similar study with BDO2 also showed A f T transversions together with deletions (38). These results are in general agreement with those reported by Cochrane and Skopek (23-25). However, sequence effects may be important; one inhalation study with butadiene has shown preferential A f G transitions in the H-ras codon 61 (41). In efforts to establish the structural basis for the various types of DNA damage that have been observed in cells treated with BD and its epoxides, mutagenesis studies have been carried out using DNA into which structurally defined adducts have been placed in a site specific manner (42-47). Monoadducts at N2 were strongly blocking with less than 1% mutagenic frequency (44). The mutagenicity of intrastrand BDO2 cross-links between the N2 positions of adjacent guanines has been examined; the lesion resulted in severely decreased plaque forming ability but to the extent that replication occurred caused both point mutations and deletions (42). Point mutations at AT sites are of interest even though they are less prevalent than mutations at GC sites. The N6 adducts of BDO were nonmutagenic in bacteria and those of BDE (or BDO2 after hydrolysis of the second epoxide) showed only a low level of mutagenicity (43). However, a study of the N1 deoxyinosine adduct of BDO,

1574

Chem. Res. Toxicol., Vol. 15, No. 12, 2002

also carried out in bacteria, showed it to be highly mutagenic, causing a predominance of A f G mutations (45). Nevertheless, this observation must be contrasted with the A f T mutations that predominated in Recio’s studies (30-40). In the present investigation, the search has been continued for the basis of the mutations that are observed at AT sites in mammalian systems. With other types of DNA adducts, several studies have shown that mammalian cells process DNA adducts differently from bacteria; in fact, some adducts can be more mutagenic in mammalian cells than in bacteria (48, 49). Consequently, the earlier studies of N1 inosine adducts of BDO which had been carried out in Escherichia coli have been repeated in simian kidney cells (COS-7). In addition, the mutagenicity of the previously unstudied BDO2-derived intrastrand cross-links (Scheme 2) between the N6 positions of adjacent deoxyadenosines has been examined in both the mammalian and bacterial system.

Materials and Methods Materials. T4 DNA ligase, T4 polynucleotide kinase, and EcoRV were obtained from New England BioLabs (Beverly, MA). S1 nuclease was purchased from Life Technologies, Inc. (Rockville, MD). [γ32P]ribo-ATP was purchased from NEN Life Science Products. Inc. (Boston, MA). Bio-Spin columns were purchased from Bio-Rad (Hercules, CA). Centricon 100 concentrators were obtained from Amicon, Inc. (Beverly, MA). Dulbecco’s Modified Eagle Medium (DMEM), fetal bovine serum, Opti-MEM (reduced serum medium), L-glutamine, antibiotic-antimycotic, and lipofectin reagent for tissue culture were obtained from Life Technologies, Inc. (Rockville, MD). Trypsin-EDTA and HEPES buffer were purchased from Cellgro Mediatech (Herndon, VA). Phosphate-buffered saline was from Sigma Chemical Co. (St. Louis, MO). Oligodeoxynucleotides containing the two diastereomers of the N1-BDO adducts of deoxyinosine were prepared as previously described using the phosphoramidite reagent of N1-(1acetoxy-3-buten-2-yl)deoxyinosine and purified by the DNA Synthesis Core Laboratory at Vanderbilt University as previously described (50). An unmodified 11-mer with adenine in place of the adducted/ cross-linked adenine was purchased from Midland Certified Reagent Co. (Midland, TX). All other unmodified oligodeoxynucleotides were synthesized by the Molecular Biology Core Laboratory at the University of Texas Medical Branch and purified by electrophoresis through a 15% denaturing PAGE in the presence of 7 M urea. DNA Purification. In the synthesis of modified oligodeoxynucleotides, HPLC purifications were carried out on a Beckman HPLC (System Gold software, model 125 pump, model 168 photodiode array detector). Oligodeoxynucleotides were desalted on Sephadex G-25 using a Bio-Rad FPLC system. 1H NMR spectra were recorded at 300.13 and 400.13 MHz on Bruker AC300 and AM400 NMR spectrometers in MeOD-d4 or DMSOd6. High-resolution mass spectra were obtained in positive ion, fast-atom bombardment (FAB) mode at the Mass Spectrometry Facility at the University of Notre Dame, Notre Dame, Indiana. Mass spectra (MALDI-TOF) of oligodeoxynucleotides were obtained using a Voyager Elite DE instrument (PerSeptive Biosystems). The system was operated in the negative ion mode using a matrix mixture of 2′,4′,6′-trihydroxyacetophenone monohydrate and ammonium hydrogen citrate or ammonium tartrate. For the mutagenesis studies, single stranded (ss) pMS2 DNA was obtained as a generous gift from Dr. M. Moriya (State University of New York, Stony Brook, NY). COS-7 cells were purchased from American Type Culture Collection (Rockville, MD). E. coli DH10B cells that were used for transformation of

Kanuri et al. DNA isolated from COS-7 cells after transfection with modified and control DNA, and helper phage M13KO7 were purchased from Life Technologies Inc. (Rockville, MD). Repair-deficient E. coli AB2480 (uvrA- recA-) was used for propagating the pMS2 shuttle vector during isolation of ss DNA by helper phage rescue using M13 KO7 (1011 pfu/mL). Purity of ss pMS2 was assayed by electrophoresis through a 1.4% agarose gel and compared with standard ss pMS2 DNA as supplied by Dr. M. Moriya. Synthesis and Purification of 11-Mer Oligodeoxynucleotides Containing dA-dA Intrastrand Cross-links. The 11mer, 5′-CGGACXXGAAG-3, where X ) 6-chloropurine, was prepared as previously described using an Expedite 8909 nucleic acid synthesizer with tert-butyl-phenoxyacetyl (tBPA) 2-cyanoethyl phosphoramidites and the phosphoramidite derivative of the 2′-deoxyriboside of 6-chloropurine on a 1-µmol scale (51). Synthesis of 2R,3R- and 2S,3S-Diastereomers of the BisNucleoside. 9-(2′-Deoxyribosyl)-6-chloropurine (38 mg, 0.14 mmol), 2S,3S- or 2R,3R-1,4-diamino-2,3-butanediol (42) (0.035 mmol), diisopropylethylamine (30 µmol), and distilled DMSO (300 µL) were mixed in a small glass test tube. The tube was flushed with argon and sealed and the reaction mixture was stirred at 60 °C. After ∼5 h, HPLC analysis of an aliquot from the reaction showed complete disappearance of the starting material. The solvents were evaporated to dryness (vacuum centrifuge), and after flash silica gel chromatography (acetonitrile:water:ammonium hydroxide, 85:8:7), 1,4-bis(2′-deoxyadenosin-N6-yl)butane-2,3-diol was isolated in 65-75% yield. The monoadducted N6-aminoalkyl deoxyadenosine was isolated as a side product (∼10%) in both syntheses. The two diastereomers of 1,4-bis(2′-deoxyadenosin-N6-yl)butane-2,3-diol had essentially identical spectra. 1H NMR (MeOD, 35 °C): δ 8.24 (s, 2H, 2 × H-8), 8.19 (s, 2H, 2 × H-2), 6.41 (dd, 2H, 2 × H-1′, J1 ) J2 ) 6.1 Hz), 4.56 (m, 2H, 2 × H-3′), 4.05 (m, 2H, 2 × H-4′), 3.89 (m, 4H, 2 × CH2N, 2 × CHO), 3.83 (m, 2H, 2 × H-5′), 3.73 (m, 4H, 2 × H-5′, 2 × CH2N), 2.79 (m, 2H, 2 × H-2′), 2.39 (m, 2H, 2 × H-2′′). HRMS (FAB+) m/z calcd for [M + H]+ 589.2482; found for the R,R-diastereomer: 589.2476, found for the S,Sdiastereomer: 589.2477. Synthesis of the BDO2-Cross-Linked Oligodeoxynucleotides. In parallel syntheses, the two diastereomers of 1,4diamino-2,3-butanediol (2.5 mg) were dissolved in borate buffer (250 µL 0.05 M Na2B4O7-NaOH, pH 10) and aliquots of these solutions containing 0.15 µmol of diamine were added to solutions of the chloropurine-containing oligodeoxynucleotide (40 A260 units, ∼0.3 µmol) in the pH 10 borate buffer (200 µL). The reactions were stirred for 24 h at 60 °C. Additional diamine solution (equivalent to 0.09 µmol of diamine) was added, and the reactions were stirred at 60 °C for another 2 days. The reactions were monitored by reversed-phase HPLC (YMC ODSAQ, 4.5 × 250 mm), with the following gradient: (A) 0.1 M ammonium formate and (B) CH3CN, 1 to 10% B over 15 min, 10 to 20% B over 5 min, at a flow rate of 1.5 mL/min. The starting oligodeoxynucleotide eluted at 17.6 min; the BDO2cross-linked oligodeoxynucleotides eluted at 14.6 min. Preparative HPLC was carried out on a 10.0 × 250 mm YMC-ODS-AQ column at 5 mL/min with the following gradient: (A) 0.1 M ammonium formate and (B) CH3CN, 1 to 5% B over 3 min, 5 to 7% B over 30 min, 7 to 80% B over 4 min. The BDO2-crosslinked oligodeoxynucleotides eluted at 37 min. The reactions were stopped at 90 h. Yields were 25-30%, determined by UV spectroscopy at 254 nm, assuming the molar absorptivity of the adducted deoxyadenosines was the same as unadducted deoxyadenosine. All products were characterized by their MALDITOF spectra and a comparison of enzyme digests with authentic samples of the nucleosides. Digestion of the cross-linked oligodeoxynucleotides was incomplete. Consequently, additional characterization was carried out by depurinating the crosslinked oligodeoxynucleotide with 0.1 M HCl (70 °C for 1 h) followed by HPLC comparison of the bases released by the acid treatment and by comparable acid treatment of the cross-linked bisnucleosides. Oligodeoxynucleotide containing the S,S-BDO2 linker: MALDI-TOF MS [M - H]-: calcd 3483.7; found 3483.9.

Mutagenesis of Butadiene-Induced Adducts

Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1575

Figure 1. Construction of circular ss pMS2 DNA modified with butadiene adducts. Single-stranded pMS2 shuttle vector suitable for replication in mammalian cells was annealed to a complementary 57-mer scaffolding DNA to stabilize the EcoRV site hairpin loop digested with EcoRV, and hybridized to a 11-mer nondamaged control or a butadiene adduct/cross-link containing single stranded oligodeoxynucleotide, and ligated to recircularize the ss DNA. UDG-Uracil DNA glycosylase. Oligodeoxynucleotide containing the R,R-BDO2 linker: MALDITOF MS [M - H]-: calcd 3483.7; found 3483.2. Construction of Circular ss pMS2 DNA Modified with Butadiene Adducts. ss pMS2 (29 pmols, 50 µg) was annealed to a 57-mer scaffold (145 pmols/2.34 µg) in the presence of buffer A containing 100 mM NaCl, 50 mM Tris-HCl, 10 mM MgCl2, 1 mM dithiothreitol (DTT), and 100 µg/mL of bovine serum albumin (BSA), in 200 µL of reaction mixture, and incubated overnight at 9 °C (Figure 1). The hairpin loop in the ss pMS2 DNA that contains a single EcoRV site was digested with 40 units of EcoRV in the presence of buffer A in a 600 µL reaction mixture, by incubation for 3 h at 37 °C (Figure 1). The 57-mer scaffold was complementary to the two termini of the digested vector (after removal of the hairpin loop), and the central sequence was complementary to the 11-mer oligodeoxynucleotide bearing the butadiene adducts, except that there was complete substitution of uracil for thymine. Nonannealed 57-

mer scaffold and the excised hairpin loop were removed from the reaction by centrifugation in a Centricon-100 tube. The gap created by annealing of ss pMS2 DNA to the scaffold was filled by ligation with a 10-fold molar excess of the adduct containing oligodeoxynucleotide, or the unmodified 11-mer as control. The 11-mer oligodeoxynucleotides were phosphorylated at the 5′ end with ribo-ATP and ligated in a buffer containing 50 mM TrisHCl (pH 7.5), 10 mM MgCl2, 10 mM DTT, 10 mM ATP, and 25 µg/mL of BSA using T4 polynucleotide kinase (5 units/pmol) and T4 ligase (325 units/pmol) in the reaction mixture of 200 µL for 48 h at 4 °C. Ligation products were visualized by ethidium bromide staining following electrophoresis through a 1.4% agarose gel and compared with standard circular and linearized ss pMS2 DNA. Modified samples were purified to remove nonligated 11mer in Centricon-100 tubes after dilution to 600 µL. After repeating this step three times, DNA was recovered by ethanol

1576

Chem. Res. Toxicol., Vol. 15, No. 12, 2002

precipitation and the DNA concentration determined by measuring absorbance at 260 nm. To avoid the 57-mer scaffold being used as a primer for (-) strand replication, ligated DNAs were incubated with uracil-DNA glycosylase for 1 h to remove uracil bases and generate abasic sites. Transfection of Modified DNA into COS-7 Cells. Ligated DNAs were transfected into COS-7 cells (5 × 105 cells/6 cm Petri dish), that had been grown in DMEM (supplemented with 10% fetal calf serum, 100 units/mL penicillin, and 100 µg/mL streptomycin) and incubated for 24 h in 5% CO2. Individually, 30 µg of lipofectin and 200 ng of ligation mixture were added to in 1.5 mL of Opti-MEM and incubated for 15 min at room temperature. A total of 3 mL of this solution was added to each 6 cm Petri plate containing the attached COS-7 cells and incubated for 18 h at 37 °C. Transfection medium was then replaced with 5 mL of fresh DMEM and the cells cultured for an additional 48 h. Cells were then washed twice with 5 mL of phosphate-buffered saline, and progeny phagemid DNA was recovered by the method described by Hirt (52). Site-Specific Mutagenesis Assay. DNA extracted from COS-7 cells was successively treated with S1 nuclease (0.2 U/60 µL reaction mix) and EcoRV (20 units/60 µL reaction mix) in order to remove any input ssDNA and progeny associated with the original ss pMS2 vector and then replicated in E. coli DH10B. A total of 50 µL of electrocompetent DH10B cells was transformed with 5 µg of DNA by electroporation in chilled 2-mm gap cuvettes using ECM 600 (BTX, San Diego), at 25 µF, 2.5 kV, R5 (Resistance timing) 129 ohms. 2 × yeast-tryptone (YT) (0.95 mL) was immediately added and incubated with shaking at 37 °C for 1 h to fully express the ampicillin-resistant gene. Aliquots of the transformation mixture were plated onto 1 × YT agar plates containing 100 µg/mL of ampicillin. After incubation overnight at 37 °C, the number of transformants per plate was counted to determine replication efficiency. In addition, 5 µg of unadducted and adduct-modified ss pMS2 DNAs were added to 40 µL of competent cells of either (a) E. coli AB2480 (uvrA- recA-), or (b) E. coli AB1157 (wild-type) cells, or (c) AB1157 cells subjected to UV irradiation at 20 J/m2 in order to induce SOS response. Transformation was achieved by electroporation under the conditions described above, and the cells were allowed to replicate on LB-ampicillin plates overnight, for comparison of mutagenicity in mammalian and bacterial cells. A total of 480 colonies from each plate of transformed cells were individually picked with sterile toothpicks and grown overnight in 2 × YT containing ampicillin in 96-well plates. After overnight growth, the cultures from each plate were transferred, using 48-pin replicators, onto Whatman 541 filter papers placed on ampicillin containing 1 × YT agar plates, in four replicates, one for each 17-mer probe (5′-GATCGGACNAGAAGATC-3′, where N refers to A, T, G, or C), and incubated overnight at 37 °C to form well-defined colonies. Alternatively, colony lifts were performed on four replicate plates of transformants in order to determine the actual mutation frequency of the entire population. Filter papers were removed from the agar plates, cell lysis achieved by soaking in 0.5 M NaOH for 5 min followed by neutralization in 0.5 M Tris-HCl, (pH 7.4) and washed in 1 × SSC (3 M NaCl; 0.3 M C6H5O7Na3, pH 7.0). DNA was cross-linked onto the paper in a UV Stratalinker for 1.5 min. Hybridization was carried out with [γ32P]ATP labeled probes, designed to determine the nature of the base at the site of the adducted adenine/inosine in progeny plasmid DNA. Hybridization temperature was calculated using the formula where the oligodeoxynucleotide melting temperature equals [4(G + C) + 2(A + T)] - 14 °C. Filters were incubated in prehybridization solution containing 25 mL of 20 × SSPE (3 M NaCl; 200 mM NaH2PO4; 20 mM EDTA), 20 mL of formamide, 5 mL of Denhardt’s solution [1% Ficoll; 1% poly(vinylpyrrolidone); 1% BSA], 1 mL of 10 mg/mL fish milt DNA, and 1 mL of 10% SDS for 20 min to restrict nonspecific binding of the probe. Four filters were separated and probed with 100 pmol each of four

Kanuri et al. 17-mer oligodeoxynucleotides containing one of the four canonical bases at the adducted site within the centrally located sequence of the damaged and nondamaged 11-mer oligodeoxynucleotide. After overnight hybridization with probes in hybridization solution, filters were washed with 2 × SSPE at 2 °C above hybridization temperature to remove nonspecifically bound probes. Dried filters were exposed to X-OMAT AR film overnight and autoradiographs were developed to identify frequency and types of mutations. Representative colonies were subjected to dideoxy sequencing (53) to confirm the sequence of the mutated DNAs. A 20-mer primer (5′-CCATCTTGTTCAATCATGCG-3′) was used for sequencing progeny plasmid DNA.

Results Construction of the Oligodeoxynucleotides Containing the Butadiene Lesions (Scheme 2). The initial reaction of BDO2 with adenine appears to be primarily at N1 (10) leading to a N1-N1 cross-link (Scheme 2) if a second adenine were appropriately situated. This cross-link would be susceptible to Dimroth rearrangement and deamination. The cross-linked adducts used in the present study are those that would arise from Dimroth rearrangement of products initially formed by reaction of R,R- or S,S-BDO2 at N1 of adenine. The 11-mer oligodeoxynucleotides containing the N6 dA-dA cross-links of R,R- and S,S-BDO2 were prepared by treatment of an oligodeoxynucleotide which had adjacent 6-chloropurine residues at the positions 61.2 and 61.3 in the N-ras codon 61 sequence with R,R- and S,S1,4-diaminobutane-2,3-diols. The stoichiometry was carefully controlled to avoid reaction of two molecules of diamine with the oligodeoxynucleotide or vice versa. Reactions were run with monitoring by HPLC until the starting oligodeoxynucleotide and un-cross-linked, adducted oligodeoxynucleotide had reached negligible levels. Formation of the cross-link was facile, indicating that it was readily accommodated in the oligodeoxynucleotide structure. The 11-mers containing the N1 deoxyinosine adducts of BDO were prepared using a phosphoramidite reagent of N1-(1-acetoxy-3-buten-2-yl)deoxyinosine. Mutagenesis of BDO2-Induced N6-N6-AdenineAdenine Cross-Links in COS-7 Cells. One of the limitations in using the pMS2 shuttle vector is that individual colonies must be hand-picked into 96-well microtiter plates and analyzed in quadruplicate by differential hybridization. To increase the total number of E. coli transformants that could be assayed from DNAs replicated in COS-7 cells, direct lifts were also performed on all E. coli colonies obtained after selection for ampicillin resistance. These transformants were subjected to differential hybridization to obtain mutation frequencies for all lesions described. Since all the transformed colonies were transferred to the filter papers for mutational analyses by differential hybridization, it was not possible to subsequently sequence these DNAs to confirm the mutations; however, mutation frequencies were confirmed and found to be in excellent agreement with data obtained from hand-picked colonies. Data presented (Tables 1 and 2) are representative of all transformed colonies lifted onto filter papers and subjected to differential hybridization. The mutagenic frequencies and spectra for butadieneinduced N6-N6-adenine-adenine cross-links are presented in Table 1. As predicted, the ss pMS2 DNA containing the nondamaged 11-mer oligodeoxynucleotide, hybridized

Mutagenesis of Butadiene-Induced Adducts

Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1577

Table 1. Mutagenic Spectrum of Butadiene Cross-Links pattern of hybridization A sample nondamaged R,R-butadiene S,S-butadiene

T

G

C

cells

no.

(%)

no.

(%)

no.

(%)

no.

(%)

total mutation %

COS-7 E. coli (AB2480) COS-7 E. coli (AB2480) COS-7 E. coli (AB2480)

250 1890 200 2000 200 1248

(100) (100) (46) (92) (80) (97)

0 0 20 11 14 5

(0) (0) (5) (0.5) (5.6) (0.4)

0 0 170 150 32 30

(0) (0) (40) (7.5) (13) (2.3)

0 0 39 1 2 1

(0) (0) (9) (0.04) (0.8) (0.08)

0 0 54 8 19.4 2.8

Table 2. Mutagenic Spectrum of Deoxyinosine Adducts pattern of hybridization A

T

G

C

sample

cells

no.

(%)

no.

(%)

no.

(%)

no.

(%)

total mutation %

nondamaged

COS-7 E. coli (AB2480) COS-7 E. coli (AB2480) COS-7 E. coli (AB2480)

250 1890 112 20 880 1572

(100) (100) (5) (4) (41) (40)

0 0 120 12 85 40

(0) (0) (6) (2) (4) (1)

0 0 1714 496 1020 1690

(0) (0) (79) (87) (48) (43)

0 0 214 45 144 1

(0) (0) (10) (8) (7) (0.08)

0 0 94.5 96.5 59 53

S-inosine R-inosine

only with the wild-type complementary probe (Table 1; Figure 2). In contrast, the DNA containing the cross-link formed from R,R-BDO2 was highly mutagenic, with all types of base substitution mutations occurring at the site of the 3′ adducted adenine in the cross-link (Table 1; Figure 3A). The most frequent mutations, based on the lift studies, were A f G transitions (40%), followed by A f C (9%) and A f T (5%) transversions. Misincorporation with all three bases opposite the site of the lesion was also found when ss pMS2 DNA containing the S,SBDO2 cross-link was assayed, but at a significantly lower overall mutagenic frequency (20%) (Table 1; Figure 3C). Again, all mutations were generated opposite the 3′ adenine of the cross-link with the predominant mutation being an A f G transition at ∼13%, with A f T and A f C transversions observed at ∼6 and ∼1%, respectively. Mutagenesis of BDO-Induced Deoxyinosine Adducts. Replication of the ss pMS2 vector containing the R- and S-diastereomers of BDO-deoxyinosine adducts resulted in 59 and 98% mutation frequencies, respectively with A f G mutations occurring most frequently (Table 2; Figure 4). 80% A f G mutations, 10% A f C mutations and 8% A f T mutations were detected for the S-inosine

Figure 2. BDO2 cross-links-differential hybridization of colonies grown in 96 well plates. Autoradiogram showing spectrum of mutations in control nondamaged, R,R- and S,S-butadiene cross-link containing DNAs. A, T, G, and C denote the base present at the site of the adducted adenine at the 3′ of the 11mer oligodeoxynucleotide.

Figure 3. Mutation percentage of site-specifically modified ss pMS2 DNAs containing R,R- and S,S-butadiene cross-links. (A) Mutations in R,R-butadiene replicated in COS-7 cells. (B) Mutations in R,R-butadiene replicated in E. coli cells. (C) Mutations in S,S-butadiene replicated in COS-7 cells. (D) Mutations in S,S-butadiene replicated in E. coli cells.

Figure 4. BDO-inosine adducts-differential hybridization of colonies grown in 96-well plates. Autoradiogram showing spectrum of mutations in control nondamaged, R-inosine and S-inosine-containing DNAs. A, T, G, and C denote the base present at the site of the adducted adenine at the 3′ of the 11mer oligodeoxynucleotide.

1578

Chem. Res. Toxicol., Vol. 15, No. 12, 2002

Figure 5. Mutation percentage of site-specifically modified ss pMS2 DNAs containing R- and S-inosine adducts. (A) Mutations in R-inosine replicated in COS-7 cells. (B) Mutations in Rinosine replicated in E. coli cells. (C) Mutations in S-inosine replicated in COS-7 cells. (D) Mutations in S-inosine replicated in E. coli cells.

adduct (Table 2; Figure 5A). Results obtained from colony lifts (95%) for the S-inosine lesion were also in agreement with data obtained from a subset of colonies picked and grown in 96-well plates (98%) (Table 2). The R-BDOdeoxyinosine lesion generated fewer mutations (59%) than the S-inosine lesion (Table 2; Figure 5C). A f G transitions predominated with the R-BDO-deoxyinosine lesion also, at a mutation frequency of 48%. A f C mutations were found at a frequency of 7% and A f T mutations were at 4%. Mutation percentages for colonies grown in 96-well plates were 53% while the data for colony lifts were 59%, with the spectrum of mutations being the same for both experimental procedures. Consistent with the data described above, the undamaged control 11-mer oligodeoxynucleotide did not generate any mutations (Table 2; Figure 4), hybridizing entirely with the wild-type complementary probe. Mutagenesis in E. coli. Often, due to cost and labor considerations, E. coli cells are used to assess the mutagenic potential of suspected mutagens and carcinogens. To determine if E. coli would yield comparable mutagenic frequencies and spectra, the ligations that had been used to transfect COS-7 cells were also used to directly transform E. coli AB2480 (uvrA- recA-). In the case of the R,R- and S,S-BDO2 dA-dA cross-links, both the spectra and frequencies significantly changed in E. coli, with the R,R adduct yielding only A f G transitions at ∼8%, while the S,S lesion produced only ∼3% total mutations, with A f G transitions predominating (Table 1; Figure 3, panels B and D, respectively).Wild-type AB1157 cells were also used in these experiments for comparison of mutagenesis. These wild-type cells were also irradiated with UV light at 20 J/m2 to induce SOS response and then transformed with modified pMS2 DNAs, both control and the four adducted DNAs. Wildtype AB1157 cells also showed similar mutagenic pattern as the AB2480 cells, with a slight reduction in mutation percentages to 5 and 1.5%, respectively for R,R- and S,Sbutadiene cross-links, respectively. However, SOS-induction resulted in a 3-5-fold increase in the total number of transformants, with a further reduction in mutation percentages to 3 and