Keynote address: exocyclic adducts as new risk markers for DNA damage in man

Background levels of exocyclic DNA adducts detected by ultrasensitive methods in tissues from unexposed humans and rodents arise from endogenous lipid peroxidation products such as trans-4-hydroxy-2-nonenal, crotonaldehyde and malondialdehyde. The levels of DNA adducts in rodent and human tissues and leukocytes were found to be highly variable and to be affected by lifestyle, the dietary intake of antioxidants and the type and amount of fatty acids and persistent chronic infections or inflammations, in which nitric oxide is often over-produced. Limited evidence suggests that etheno-DNA adducts play a role not only in vinyl chloride- and urethane-induced tumorigenesis but, together with other exocyclic lesions, also in several human cancers in which persistent oxidative stress leads to malignancy by increasing mutation rates and genomic instability. Therefore, promutagenic exocyclic adducts appear to be promising markers in molecular epidemiological studies for identifying endogenous sources of DNA damage and resulting oxidative modifications in cancers with poorly defined etiology and mechanisms and in intervention studies to assess the protective effects of antioxidants against cancer and, possibly, neurodegenerative diseases.     

Potential role of lipid peroxidation derived DNA damage in human colon carcinogenesis: studies on exocyclic base adducts as stable oxidative stress markers

Molecular pathways to colorectal cancer involve multiple genetic changes that may be caused by overproduction of reactive oxygen species in cancer-related genes. Our aim was to investigate, whether besides direct oxidative DNA damage, reactive oxygen and nitrogen species induce lipid peroxidation (LPO) that could yield etheno-DNA adducts via trans-4-hydroxy-2-nonenal, a major aldehyde generated by LPO, in colon tissue. We analyzed the etheno-DNA adducts by a highly specific, ultrasensitive method involving immunoaffinity chromatography coupled with 32P-postlabelling [Carcinogenesis 16 (1995) 613] in affected colon epithelium from ulcerative colitis, Crohn's disease and familial adenomatous polyposis (FAP) and compared them with asymptomatic colon tissue. In all these cancer prone colon tissues, the formation of markedly enhanced etheno adduct levels was demonstrated for the first time. Etheno-DNA adducts are promutagenic and cause genomic instability that could drive the inflamed colonic epithelia to malignancy. Etheno-DNA adducts appear promising biomarkers for (i) quantifying increased DNA damage in early stages of colon carcinogenesis and for (ii) verifying the efficacy of new antioxidants (e.g. [Lancet Oncol. 1 (2000) 107]) and chemopreventive agents in lowering oxidative stress and related cancer risk.     

Lipid peroxidation dominates the chemistry of DNA adduct formation in a mouse model of inflammation

In an effort to define the prevalent DNA damage chemistry-associated chronic inflammation, we have quantified 12 DNA damage products in tissues from the SJL mouse model of nitric oxide (NO) overproduction. Using liquid chromatography-mass spectrometry/MS and immunoblot techniques, we analyzed spleen, liver and kidney from RcsX-stimulated and control mice for the level of the following adducts: the DNA oxidation products 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxodG), guanidinohydantoin (Gh), oxazolone (Ox); 5-guanidino-4-nitroimidazole (NitroIm); spiroiminodihydantoin (Sp) and M(1)dG; the nitrosative deamination products 2'-deoxyxanthosine, 2'-deoxyoxanosine (dO), 2'-deoxyinosine and 2'-deoxyuridine and the lipid peroxidation-derived adducts 1,N(6)-etheno-deoxyadenosine and 1,N(2)-etheno-deoxyguanosine. The levels of dO, Gh, Ox, NitroIm and Sp were all below a detection limit of approximately 1 lesion per 10(7) bases. Whereas there were only modest increases in the spleens of RcsX-treated compared with control mice for the nucleobase deamination products (10-30%) and the DNA oxidation products 8-oxodG (10%) and M(1)dG (50%), there were large (3- to 4-fold) increases in the levels of 1,N(6)-etheno-deoxyadenosine and 1,N(2)-etheno-deoxyguanosine. Similar results were obtained with the liver and with an organ not considered to be a target for inflammation in the SJL mouse, the kidney. This latter observation suggests that oxidative and nitrosative stresses associated with inflammation can affect tissues at a distance from the activated macrophages responsible for NO overproduction during chronic inflammation. These results reveal the complexity of NO chemistry in vivo and support an important role for lipids in the pathophysiology of inflammation.     

A 32P-postlabeling method for simultaneous detection and quantification of exocyclic etheno and propano adducts in DNA

A ³²P-postlabeling method is described that specifically detectsand quantifies the 1,N² adducts derived from acrolein (AdG)and crotonaldehyde (CdG) and 1,N²-ethenodexoxyguanosine (EdG)in DNA. These exocyclic adducts are potential DNA lesions causedby exposure to enals as environmental pollutants and as endogenouscompounds. This method was developed with the use of the syntheticadduct standards of these exocyclic adducts. The assay relieson HPLC for adduct enrichment prior to labeling and for quantitationand identification after labeling. The labeling efficienciesof adducts at the 1 fmol level ranged from 74 to 96%, whereasthey were only 49–60% at the 100 fmol level. This methodcan detect as low as 0.2 fmol of adduct and allows the detectionand quantitative determination of stereolsomers of AdG and CdG.The method was validated by using a sample of enzyme digestsof 180 µg calf thymus DNA spiked with 25 or 75 fmol ofadducts, which is equivalent to 5 or 15 adducts in 10⁸ nucleotides.The recovery rates of these adducts in DNA ranged from 30 to90% at the 25 fmol level and 21 to 55% at the 75 fmol level.Similar to the labeling efficiency, a greater recovery was observedwith a lower amount of adduct in DNA. Overall, this method allowsthe simultaneous identification and quantification of exocycicadducts AdG, CdG and EdG in DNA. Therefore, it provides a potentialtool for studies of the in vivo formation of exocyclic adducts.     

Role of 1,N2-propanodeoxyguanosine adducts as endogenous DNA lesions in rodents and humans

Results obtained in a number of studies in vitro and in vivo support the hypothesis that short- and long-chain enals and their epoxides derived from oxidized polyunsaturated fatty acids are potential endogenous sources of cyclic propano and etheno DNA adducts. We previously reviewed the evidence from some of these studies. Here, we describe the results of our more recent studies on the role of 1,N2-propanodeoxyguanosine adducts as endogenous DNA lesions. These studies include: the detection of distinct patterns of such adducts in various tissues of different species; the detection of long-chain trans-4-hydroxynonenal-derived deoxyguanosine adducts in vivo; the specificity of the formation of enal-derived propano adducts from omega-3 and omega-6 polyunsaturated fatty acids; and the detection of acrolein- and crotonaldehyde-derived adducts in human oral tissue DNA and their increased levels in smokers. Taken together, these studies further strengthen the hypothesis that enals produced by lipid peroxidation are the primary source for cyclic propano adducts in vivo, but these results cannot rule out the possible contribution of environmental and other sources. The mutagenicity of enals and their epoxides and the results of site-specific mutagenesis studies indicate that the cyclic adducts are potential promutagenic lesions; however, only circumstantial evidence is currently available for their role in carcinogenesis.     

In vitro formation of DNA adducts with bile acids

The in vitro experiment was carried out following ³²P-postlabelinganalysis to determine the DNA-reactive bile acids present inthe human body. The unconjugated and conjugated forms of cholic(CA), chenodeoxycholic (CDCA), deoxycholic (DCA) and lithocholicacid (LCA) were added to calf thymus DNA followed by 1 h ofincubation at 37°C. After the incubation the mixture wasanalyzed by the nuclease P1 modification of ³²P-postlabeling.Among the 12 bile acids tested, our results showed that unconjugatedCDCA and LCA and the glycine and taurine conjugates of LCA (g-LCA,t-LCA) were able to bind covalently with naked DNA in vitrowithout intervention of any catalyst. It was also shown thatbile acid alone did not give any spot on TLC. These bindingreactions depended on the bile acid concentration in a linearmanner. The data on the extent of binding at a concentrationof 0.1 mg/ml showed values of 28.5 (t-LCA), 23.7 (g-LCA), 3.47(LCA) and 1.32 (CDCA) adducts per 10⁸ nucleotides. These reactivebile acids were also incubated with COLO 205 human colon carcinomacells and Hep G2 human hepatocellular carcinoma cells for 24h, but no specific DNA adduct was formed. Further, when LCAor CDCA was administered into male Fischer 344 rats by gavageat a dose of 10 mg/rat every 8 h for 3 days, no specific DNAadduct was detected in their liver or colon. Covalent DNA adductsare believed to cause alteration of the primary structure ofgenes, which is potentially linked to carcinogenesis. Thoughour preliminary data failed to detect any bile acid-relatedDNA adducts in the cultured cells or in rats, the results mayprovide a basis for assuming some of these bile acids to bepotential initiators of colon cancer.     

Detection of exocyclic guanine adducts in hydrolysates of hepatic DNA of rats treated with N-nitrosopyrrolidine and in calf thymus DNA reacted with alpha-acetoxy-N-nitrosopyrrolidine

This report describes the isolation and characterization of DNA adducts formed in vitro from alpha-acetoxy-N-nitrosopyrrolidine and in rats treated with the hepatocarcinogen N-nitrosopyrrolidine. Esterase-catalyzed hydrolysis of alpha-acetoxy-N-nitrosopyrrolidine in the presence of calf thymus DNA, followed by neutral thermal hydrolysis of the DNA, resulted in formation of three previously unknown Adducts 1-3. They were isolated and characterized by their UV, mass, and proton magnetic resonance spectra as the exocyclic 7,8-guanine adducts 2-amino-6,7,8,9-tetrahydro-9-hydroxypyrido[2,1-f]purine-4(3H)-one (Adduct 1), and cis- and trans-2-amino-7,8-dihydro-8-hydroxy-6-methyl-3H-pyrrolo[2,1-f] purine-4(6H)-one (Adducts 2 and 3). Adduct 1 was formed by addition of 4-oxobutyl diazohydroxide, or a related carbonium ion, to the 7 and 8 positions of guanine. Adducts 2 and 3 resulted from Michael addition of 2-butenal to the 7 and 8 positions of guanine. Esterase-catalyzed hydrolysis of alpha-acetoxy-N-nitrosopyrrolidine in the presence of DNA also produced the exocyclic 1,N2-propanodeoxyguanosine Adducts 4a and 4b which we have previously described. Neutral thermal hydrolysates of hepatic DNA isolated from rats treated with N-nitrosopyrrolidine contained a fluorescent adduct, as previously reported (E.J. Hunt and R.C. Shank, Biochem. Biophys. Res. Commun., 104: 1343, 1982). This fluorescent adduct was shown to be identical to Adduct 1. Adducts 2, 3, 4a, and 4b were not detected in hepatic DNA hydrolysates from these animals. The results of this study provide the first example of a structurally characterized DNA adduct formed in vivo from a cyclic nitrosamine and support the alpha-hydroxylation hypothesis of cyclic nitrosamine metabolic activation.     

Lipid peroxidation and DNA adduct formation in lymphocytes of premenopausal women: Role of estrogen metabolites and fatty acid intake

A diet high in linoleic acid (an ω-6 PUFA) increased the formation of miscoding etheno (ε) - DNA adducts in WBC-DNA of women, but not in men (Nair et al., Cancer Epidemiol Biomark Prev 1997;6:597-601). This gender specificity could result from an interaction between ω-6 PUFA intake and estrogen catabolism, via redox-cycling of 4-hydroxyestradiol (4-OH-E(2) ) and subsequent lipid peroxidation (LPO). In this study, we investigated whether in premenopausal women LPO-derived adducts in WBC-DNA are affected by serum concentration of 2- and 4-hydroxyestradiol metabolites and by fatty acid intake. DNA extracted from buffy coat and plasma samples, both blindly coded from healthy women (N = 124, median age 40 year) participating in the EPIC-Heidelberg cohort study were analyzed. Three LPO-derived exocyclic DNA adducts, εdA, εdC and M(1) dG were quantified by immuno-enriched (32) P-postlabelling and estradiol metabolites by ultra-sensitive GC-mass spectrometry. Mean M(1) dG levels in WBC-DNA were distinctly higher than those of εdA and εdC, and all were positively and significantly interrelated. Serum levels of 4-OH-E(2) , but not of 2-OH-E(2) , metabolites were positively related to etheno DNA adduct formation. Positive correlations existed between M(1) dG levels and linoleic acid intake or the ratios of dietary linoleic acid/oleic acid and PUFA/MUFA. Aerobic incubation of 4-OH-E(2) , arachidonic acid and calf thymus DNA yielded two to threefold higher etheno DNA adduct levels when compared with assays containing 2-OH-E(2) instead. In conclusion, this study is the first to compare M(1) dG and etheno-DNA adducts and serum estradiol metabolites in human samples in the same subjects. Our results support a novel mechanistic link between estradiol catabolism, dietary ω-6 fatty acid intake and LPO-induced DNA damage supported by an in vitro model. Similar studies in human breast epithelial tissue and on amplification of DNA-damage in breast cancer patients are in progress.     

Deoxyguanosine adducts of t-4-hydroxy-2-nonenal are endogenous DNA lesions in rodents and humans: detection and potential sources

t-4-Hydroxy-2-nonenal (HNE) is a free radical-mediated oxidation product of polyunsaturated fatty acids. As an electrophile, HNE readily binds to proteins and yields diastereomeric cyclic 1,N2-propano adducts with deoxyguanosine (dG). Here, we report the detection and identification of the HNE-derived cyclic 1,N2-propano-dG adducts as endogenous DNA lesions in tissues of untreated rats and humans using a highly sensitive 32P-postlabeling method in conjunction with high-performance liquid chromatography. These adducts were first verified by their comigration with the synthetic UV standards of HNE-dG adducts. Subsequently, their identities were unequivocally established by two independent reactions. An approximately 37-fold increase in the levels of HNE-dG adducts was observed in the liver DNA of F344 rats after treatment with CCl4, suggesting that tissue lipid peroxidation is a likely source of their formation. Our studies in vitro further indicate that omega-6 polyunsaturated fatty acids are likely a unique class of fatty acids involved in HNE-dG adduct formation.     

Thermal destabilization of DNA oligonucleotide duplexes by exocyclic adducts on adenine or cytosine depends on both the base and the size of adduct

Numerous carcinogens or their bifunctional metabolites modify DNA bases by forming additional exocyclic rings on the base moiety. These modifications form exocyclic rings between N1 and N6 of dA, N3 and N4 of dC or N1 and N2 of dG, as well as the N2 and N3 of dG. This study focuses on the reaction products of dA and dC with chloroacetaldehyde, bis-chloroethyl nitrosourea and para-benzoquinone, which form etheno, ethano and para-benzoquinone derivatives, respectively. The three dC adducts and three dA adducts were each incorporated site-specifically into 25-nucleotide-long deoxyoligonucleotides. All duplexes with a single modified dA or dC adduct opposite the normal complement showed decreased thermal stability, as compared with the unmodified control duplex. The destabilizations ranges from -2 degrees C to -13 degrees C, depending on saturation, the size of the adduct and the nature of the base. Energy-minimized molecular models of the duplexes illustrate various degrees of distortions by the adducts, the para-benzoquinone adducts showing the greatest distortion.     

Reactivity of benzo[a]pyrene-7,8-dione with DNA. Evidence for the formation of deoxyguanosine adducts

Polycyclic aromatic hydrocarbon (PAH) o-quinones are productsof the dihydrodiol dehydrogenase-catalyzed oxidation of trans-dihydrodiolswhich are proximate carcinogens. The PAH o-quinones are highlyreactive molecules and have the potential to alkylate DNA. Inthis study, the reactivity of [³H](+/–)-trans-7,8-dihydroxy-anti-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene([³H](+/–)-anti-BPDE), [³H]benzo[a]pyrene-7,8-dione ([³H]BPQ)and [³H](+/–)-trans-7,8-dihydroxy-7,8-dihydrobenzo[a]pyrene ([³H](+/–)-B[a]P-diol) with DNA were compared.(+/–)-anti-BPDE reacted equally well with native, deproteinatedand deproteinated/sheared calf thymus DNA. In each case DNAadducts were formed which upon digestion to deoxyribonucleosidescomigrated on reverse-phase (RP)-HPLC with adducts synthesizedby reacting (+/–)-anti-BPDE with oligo-p(dG)₁₀. (+/–)-anti-BPDEalso reacted with plasmid (pGEM-3) DNA to yield multiple adductsone of which comigrated with the (+)-anti-BPDE-deoxyguanosineadduct. Under identical conditions [³H]BPQ reacted preferentiallywith native calf thymus DNA but displayed low reactivity withdeproteinated and deproteinated/sheared calf thymus DNA. RP-HPLCanalysis of deoxyribonucleoside—BPQ adducts indicatedthat the predominant adduct formed comigrated with a standardsynthesized by reacting BPQ with oligo-p(dG)₁₀. BPQ also reactedwith pGEM-3 DNA to yield multiple adducts one of which comigratedwith the BPQ—deoxyguanosine adduct. Reactions between[³H]BPQ and poly(dA), poly(dT), poly(dC) and oligo-p(dG)₁₀ indicatedthat BPQ preferentially formed deoxyguanosine adducts. In thisstudy, [³H]BPQ and [³H](+/–)-anti-BPDE covalently labelednative calf thymus DNA to an equal extent, however, less [³H]BPQwas recovered as deoxyguanosine adducts. By contrast, no covalentmodification of calf thymus DNA, pGEM-3 DNA or oligonucleotideswas observed with [³H](+/–)-B[a]P-diol. These studiesindicate that BPQ has the potential to be genotoxic in vitro;that reactivity is heightened in the presence of protein orcircular DNA and that the major adduct formed is a deoxyguanosineadduct.     

Detection of dopamine DNA adducts: potential role in Parkinson's disease

Oxidation of catecholamines may lead to the formation of O-semiquinonesand o-quinones in catecholaminergic brain tissues, and thesereactive molecules may form DNA or protein adducts. In thisstudy, cultured cells were treated with dopamine (DA) for ²⁴h and ³²P-postlabeling was used to detect DA-DNA adducts. InHL–60 cells, 250µM DA induced 8.5 DNA adducts/108nucleotides; adduct formation was dose-dependent up to 500 µMDA. Addition of H2O2 increased the relative adduct levels ⁷-to 13-fold, but no adducts were detected when DA and ascorbicacid were added simultaneously. In human glioblastoma cell linesU87, U251, SF-763 and SF-767, 1000 µM DA produced 0.98–2.31adducts/108 nucleotides. These results suggest that the formationof DNA adducts by DA may contribute to the development of certainneurodegenerative diseases such as Parkinson's disease.     

2,3-Epoxy-4-hydroxynonanal as a potential tumor-initiating agent of lipid peroxidation

Trans-4-hydroxy-2-nonenal (HNE) is a product of lipid peroxidation.In the presence of t-butyl hydroperoxide the racemic HNE readilyconverts to its epoxide, 2,3-epoxy-4-hydroxynonanal (EH), asa pair of diastereomers. In this study, the potential rolesof HNE and EH as tumor initiating agents were assessed. Themutagenicities of HNE and EH isomers in Salmonella strains TA100and 104 were examined. In addition, the tumor initiating activitiesof HNE and EH were evaluated in bioassays involving either topicalapplication in CD-1 mice or i.p. administration in newborn CD-1mice. In the mutagenicity assays, EH isomers induced similarlevels of revertants in both tester strains, although EG isomerswere previously shown to react with bases in DNA with differentspecificity (Sodum,R.S. and Chung,F.-L., Cancer Res., 51,137–143,1991). The major isomer induced     

Genetic effects of exocyclic DNA adducts in vivo: heritable genetic damage in comparison with loss of heterozygosity in somatic cells

Etheno adducts are promutagenic lesions which generate point mutations, deletions, homologous recombination and gross structural DNA aberrations. High ratios of chromosome loss to forward mutations characterize vinyl bromide, vinyl chloride, ethyl carbamate, vinyl carbamate and its epoxide as effective clastogens in postmeiotic germ cells of Drosophila melanogaster. Of the mutants induced by vinyl carbamate at the vermilion gene, 68% were intra-locus or multi-locus deletions. In view of the far-reaching concordance between mutation spectra in mice and Drosophila observed in specific-locus tests with genotoxic agents, etheno bases are expected to generate mainly deletions in male mammals in the postmeiotic germ-cell stages. Twenty-two of 23 base substitutions induced in the vermilion gene after treatment of postmeiotic stages with vinyl carbamate or vinyl bromide fall into four categories of mutations expected from etheno bases: GC-->AT, AT-->GC, GC-->TA and AT-->TA base-pair changes. These types of point mutations occurred in mutated proto-oncogenes of tumours induced in rodents by vinyl chloride, ethyl carbamate or their metabolites. Of interest is the ability of vinyl carbamate to produce persistent lesions in otherwise highly repair-active premeiotic cells of Drosophila, leading to mutations of yet unknown nature. Etheno bases are also potent pro-clastogenic lesions in somatic cells in vivo. Strongly positive responses were reported for ethyl carbamate and vinyl carbamate in assays for micronucleus formation in mouse bone marrow and in the Drosophila white+/white eye mosaic test. Loss of heterozygosity in somatic cells of Drosophila was due primarily to ring-X chromosome loss, followed by homologous mitotic recombination. Particularly striking is the near failure of ethyl carbamate and vinyl carbamate to generate significant frequencies of intrachromosomal recombination. The overall genetic activity profiles of etheno adduct-forming chemicals in mice and in Drosophila support the hypothesis that vinyl carbamate is the proximate mutagen of ethyl carbamate, and vinyl carbamate epoxide is the ultimate electrophilic mutagen and carcinogen.     

Three-dimensional fluorometry for the detection of DNA adducts

We describe the interfacing of a fluorometer to a desk-top computer by means of a commercially available interface box, for the purpose of generating three-dimensional fluorescence spectra. The important features of a self-designed program in BASIC are discussed in detail.     

Metabolism of 1-nitropyrene and formation of DNA adducts in Salmonella typhimurium

1-Nitropyrene is slowly reduced by intact cells of Salmonellatyphimurium to yield 1-aminopyrene and N-acetyl-1-aminopyreneplus six unidentified minor products. When the bacteria areexposed to tritiated 1-nitropyrene, increasing amounts of radioactivitybecome bound to DNA as the nitropyrene is metabolized. Enzymatichydrolysis of the labelled DNA yields low molecular weight labelledcompounds which probably represent nucleoside adducts formedby the reaction of nitropyrene metabolites with DNA. Resultswith appropriate mutant strains indicate that bacterial nitroreductasesare involved in activating nitropyrene to a reactive intermediatethat binds to DNA and that nitropyrene adducts in DNA are subjectto excision repair.     

Detection of DNA hypermethylation as a potential biomarker for prostate cancer

Prostate cancer is one of the most common malignant diseases in men above the age of 50. A genetic predisposition and/or acquired genetic and epigenetic changes together with lifestyle contribute to the development of the disease. The most studied epigenetic modification in prostate cancer is the methylation of the cytosine located within the dinucleotide CpG of promoter regions of different genes by methylation specific PCR. The evidence of gene silencing by DNA methylation in genes like GSTP1, APC or RASF1 is a common and relatively specific event in prostate cancer. DNA methylation testing can be performed on tissue samples or urine, ejaculate or serum. Translational research is searching for new biomarkers for early detection and prognosis of prostate cancer, but because of large methodological differences in applied techniques and patient cohorts, the investigations have yielded promising, but also some controversial results. More prospective randomized trials and standardized methods are needed to assess the true value of methylation for the diagnosis and prognosis of prostate cancer.     

1,N2-ethenodeoxyguanosine as a potential marker for DNA adduct formation by trans-4-hydroxy-2-nonenal

The reaction of trans-4-hydroxy-2-nonenal, a major alpha, beta-unsaturated aldehyde released during lipid peroxidation, with deoxyguanosine under physiological conditions was investigated in order to assess its DNA damaging potential. This aldehyde was dissolved in tetrahydrofuran (THF) prior to addition to the reaction mixture. The results showed that structurally different adducts were formed in these reactions depending on the THF used. Using THF unprotected from light, reactions yielded adducts 1 to 6. Adduct 1 was characterized as 1,N2-ethenodeoxyguanosine (5,9-dihydro-9-oxo-3-beta-D-deoxyribofuranosylimidazo[1,2-alpha]pu rine) by its UV, proton nuclear magnetic resonance, and mass spectrum and by comparison to the corresponding guanosine and guanine adducts reported in the literature. The UV spectrum of adduct 4 was indicative of a substituted 1,N2-etheno derivative. Adducts 2,3,5, and 6 were essentially identical in UV spectra and appeared to be N2-substituted deoxyguanosine diastereomers. At room temperature adducts 2,3,5, and 6 were converted quantitatively to a single product at pH 10.5. This product was shown to be identical to 1,N2-ethenodeoxyguanosine (adduct 1). Analogous conversions to 1,N2-ethenoguanine were also observed for the corresponding guanine adducts. Using THF that had been protected from the light, however, the reactions of trans-4-hydroxy-2-nonenal with deoxyguanosine gave three major adducts, 7,8, and 9. These adducts possessed UV spectra similar to that of 1,N2-propanodeoxyguanosine and were not converted to 1,N2-ethenodeoxyguanosine upon base treatment. Evidence obtained suggests that adducts 1 to 6 were formed from the reaction of deoxyguanosine with the epoxide of trans-4-hydroxy-2-nonenal generated in the presence of hydroperoxide in the light unprotected THF, whereas adducts 7 to 9 were formed by direct Michael addition. Adducts 1 to 6 were formed presumably as a result of nucleophilic addition of the exo-amino of deoxyguanosine to the aldehydic group of the epoxide of trans-4-hydroxy-2-nonenal. Base treatment of these adducts facilitated subsequent cyclization and eliminations and finally gave 1,N2-ethenodeoxyguanosine. These results demonstrated that trans-4-hydroxy-2-nonenal readily forms adducts with deoxyguanosine either by direct Michael addition or via its epoxide formation. The facile conversion of some of these adducts to a single adduct suggests that 1,N2-ethenodeoxyguanosine may provide a simple and useful marker for assessing potential DNA damage by trans-4-hydroxy-2-nonenal and related alkenals associated with lipid peroxidation.