Formation of trans-4-hydroxy-2-nonenal- and other enal-derived cyclic DNA adducts from ω-3 and ω-6 polyunsaturated fatty acids and their roles in DNA repair and human p53 gene mutation

Mutation Research 531 (2003) 25–36

Review

Formation of trans-4-hydroxy-2-nonenal- and other enal-derived cyclic DNA adducts from ␻-3 and ␻-6 polyunsaturated fatty acids and their roles in DNA repair and human p53 gene mutation Fung-Lung Chung a,∗ , Jishen Pan a , Sujata Choudhury a , Rabindra Roy a , Wenwei Hu b , Moon-shong Tang b a

b

Division of Carcinogenesis and Molecular Epidemiology, Institute for Cancer Prevention, American Health Foundation Cancer Center, Valhalla, NY 10595, USA Department of Environmental Medicine, Pathology and Medicine, New York University School of Medicine, Tuxedo, NY 10987, USA Received 28 February 2003; accepted 18 July 2003

Abstract The cyclic 1,N2 -propanodeoxyguanosine adducts, derived from ␣,␤-unsaturated aldehydes or enals, including acrolein (Acr), crotonaldehyde (Cro), and trans-4-hydroxy-2-nonenal (HNE), have been detected as endogenous DNA lesions in rodent and human tissues. Collective evidence has indicated that the oxidative metabolism of polyunsaturated fatty acids (PUFAs) is an important pathway for endogenous formation of these adducts. In a recent study, we examined the specific role of different types of fatty acids, ␻-3 and ␻-6 PUFAs, in the formation of cyclic adducts of Acr, Cro, and HNE. Our studies showed that the incubation of deoxyguanosine 5 -monophosphate with ␻-3 or ␻-6 fatty acids under oxidative conditions in the presence of ferrous sulfate yielded different amounts of Acr, Cro, and HNE adducts, depending on the types of fatty acids. We observed that Acr- and Cro-dG adducts are primarily formed from ␻-3, and the adducts derived from longer chain enals, such as HNE, were detected exclusively from ␻-6 fatty acids. Acr adducts are also formed from ␻-6 fatty acids, but to a lesser extent; the yields of Acr adducts are proportional to the number of double bonds present in the PUFAs. Two previously unknown cyclic adducts, one from pentenal and the other from heptenal, were detected as products from ␻-3 and ␻-6 fatty acids, respectively. Because ␻-6 PUFAs are known to be involved in the promotion of tumorigenesis, we investigated the role of HNE adducts in p53 gene mutation by mapping the HNE binding to the human p53 gene with UvrABC nuclease and determined the formation of HNE-dG adducts in the gene. The results showed that HNE-dG adducts are preferentially formed in a sequence-specific manner at the third base of codon 249 in the p53 gene, a mutational hotspot in human cancers. The DNA repair study using plasmid DNA containing HNE-dG adducts as a substrate in HeLa cell extracts showed that HNE adducts are readily repaired, and that nucleotide excision repair appears to be a major pathway involved. Together, results of these studies provide a better understanding of the involvement of different PUFAs in DNA damage and their possible roles in tumorigenesis. © 2003 Elsevier B.V. All rights reserved. Keywords: Cyclic DNA adduct; Polyunsaturated fatty acid (PUFA); trans-4-Hydroxy-2-nonenal (HNE); p53 gene; DNA repair

Abbreviations: dG, deoxyguanosine; Acr, acrolein; Cro, crotonaldehyde; HNE, trans-4-hydroxy-2-nonenal; Pen, trans-2-pentenal; Hep, trans-2-heptenal; Hex, trans-2-hexenal; Oct, trans-2-octenal; PUFA, polyunsaturated fatty acid; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; LNA, linolenic acid; AA, arachidonic acid; LA, linoleic acid; SA, stearic acid; NPYR, N-nitrosopyrrolidine; NER, nucleotide excision repair; BER, base excision repair; GSH, glutathione ∗ Corresponding author. Tel.: +1-914-789-7161; fax: +1-914-592-6317. E-mail addresses: [email protected], [email protected] (F.-L. Chung). 0027-5107/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.mrfmmm.2003.07.001

26

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

1. Introduction The trans-4-hydroxy-2-nonenal (HNE) is a long-chain ␣,␤-unsaturated aldehydic product (enal) form by oxidation of polyunsaturated fatty acids (PUFAs). Like its lower homologs acrolein (Acr) and crotonaldehyde (Cro), it is highly reactive, readily reacting with cellular macromolecules such as protein and DNA via a Michael addition reaction [1]. Acr and Cro are also formed through lipid peroxidation, and are ubiquitous environmental pollutants generated by combustion of fuel and by smoking cigarette. Unlike Acr and Cro, however, HNE is a unique product of ␻-6 PUFAs. Enals are capable of forming cyclic adducts with DNA bases. Their reaction with guanine yields cyclic 1,N2 -propanodeoxyguanosine adducts (1,N2 -propanodG, Fig. 1). These adducts have been detected in tissues of humans and untreated rodents as background DNA lesions [2–4]. Although the formation of 1,N2 -propanodG adducts appears to be a common reaction of deoxyguanosine with enals, the rates of reaction of enals vary considerably. In general, the shorter chain enals are more reactive than the longer chain ones. Even though enals with different lengths of carbon chains from Acr to HNE, including

pentenal (Pen), hexenal (Hex), heptenal (Hep), and octenal (Oct), have been shown to be products of lipid peroxidation [5], so far only Acr, Cro, and HNE adducts have been detected in vivo as endogenous DNA lesions. In this paper, we provide a succinct account of our first discovery of cyclic adducts as background DNA lesions in rodent tissues, followed by a summary of current evidence that indicates enals generated by oxidative metabolism of PUFAs are the endogenous sources for the formation of cyclic adducts. The main focus of this paper is to present results of our recent studies on the roles of different types of fatty acids, ␻-3 versus ␻-6, in the formation of the enal-derived cyclic adducts. Finally, we discuss briefly our studies on sequence-specific binding of HNE in the human p53 gene and the repair of HNE-dG adducts in plasmid DNA by HeLa cells. 2. Detection of 1,N 2 -propanodG adducts as endogenous background lesions The detection of 1,N2 -propanodG adducts were first reported from a study in which we investigated

N

N HO

N H

N

O

OH

O

N

N

N R'

N H

N

N

dR

dR R'

Acr-dG 1&2 Acr-dG3 Cro-dG Pen-dG Hex-dG Hep-dG Oct-dG HNE-dG

dR =

deoxyribose

Fig. 1. Structures of the cyclic 1,N2 -propanodG adducts.

H CH3 CH3CH2 CH3(CH2)2 CH3(CH2)3 CH3(CH2)4 CH3(CH2)4CH(OH)

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

27

Table 1 Species and tissues shown to contain Acr-, Cro-, and HNE-dG Adduct

Species

Tissue

Adduct level (adduct/106 guanine)

Acr- and Cro-dG

Mouse Rat

0.31–1.69 0.02–1.12

Human

Liver, skin Liver, lung, colon, brain, prostate, mammary gland, kidney Liver, mammarygland, leukocytes

0.01–2.11

Rat Human

Liver, colon Colon

0.001–0.2 0.001–?

HNE-dG

DNA damage by N-nitrosoprolidine (NPYR), a cyclic nitrosamine and a potent liver carcinogen [6]. Because Cro is a metabolite of NPYR, as expected, we detected Cro-dG adducts in the liver DNA of F344 rats treated with NPYR using a PEI cellulose plate-based 32 P-postlabeling method. Cro-dG adducts were also detected in the skin DNA of mice treated topically with Cro. However, a background radioactive spot on the PEI plates after autoradiography was consistently detected in the liver DNA of the untreated rats, which co-migrated with the synthetic standards of Cro-dG adducts. This spot represented approximately 25% of the adducts detected in the NPYR-treated rat liver DNA, based on radioactivity. One fortuitous discovery provided the first evidence of an enal-derived cyclic dG adduct as a background DNA lesion in vivo. This observation was later confirmed by using a 32 P-postlabeling/HPLC method developed specifically for the detection and quantification of Acr- and Cro-dG adducts [2]. Subsequently, a similar method for the detection of HNE-dG adducts was developed [4]. Using these highly sensitive and specific methods, we have confirmed the role of Acr, Cro, and HNE adducts as endogenous DNA lesions in rodent and human tissues. Table 1 summarizes the levels of Acr-, Cro- and HNE-dG adducts detected in DNA from various tissue sources.

3. Lipid peroxidation is an endogenous pathway for the enal-derived cyclic adducts Because oxidative metabolism of PUFAs is known to produce a host of enals with different chain lengths [5], it was speculated that lipid peroxidation may be

the origin of the formation of the cyclic adducts in vivo. Data from in vitro and in vivo studies support the possibility that enals derived from oxidation of PUFAs are indeed an important source of cyclic adduct formation. The evidence is outlined as follows: (1) the formation of the cyclic DNA adducts is increased in the livers of rats administered CCl4 or in the livers of Long Evans Cinnamon rats which are afflicted with increased lipid peroxidation as a result of genetically predisposed copper accumulation in that tissue [7,8]; (2) studies showed that depletion of GSH in the liver of rats treated with buthione sulfoximine results in a significant increase of 1,N2 -propanodG adducts [9]; (3) a higher level of cyclic adducts occur in tissues of high fat content, such as brain, liver, and colon and in lymphocyte DNA of individuals with high fatty intake [3,10,11]; (4) the detection of HNE-dG adducts in these tissues provides further support that the cyclic adducts derived from enals are originated from oxidative reaction of PUFAs [4]; and (5) studies in vitro show that the cyclic DNA adducts are formed upon incubation with PUFAs under oxidative conditions [4,12]. 4. The role of ␻-3 versus ␻-6 PUFAs in the formation of the cyclic 1,N 2 -propanodG adducts The roles of ␻-3 and ␻-6 fatty acids in cancer development have been extensively investigated in animal and epidemiological studies. The contrasting roles of the fatty acids, the former being protective and the latter promoting, in tumorigenesis, have been an intriguing question in the field of nutrition and cancer, as the mechanisms for their differences are presently

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

(A)

O N

HN N H

HO

UV254nm

unclear. We recently examined the role of ␻-3 and ␻-6 fatty acids in the formation of 1,N2 -propanodG adducts [13]. The fatty acids studied include docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA), and linolenic acid (LNA) as ␻-3 PUFAs, arachidonic acid (AA) and linoleic acid (LA) as ␻-6 PUFAs; and stearic acid (SA) as a saturated fatty acid. Fatty acids were incubated with dG 5 -monophosphate at 37 ◦ C in the presence of Fe(II). To identify the products from the reactions, a number of adduct standards were prepared, including Acr-, Cro-, Pen-, Hex-, Hep-, Oct-, and HNE-dG 5 -monophosphates. The reaction mixtures were analyzed and the reaction products were purified by a series of reverse-phase HPLC systems by collecting the fractions corresponding to the 5 -monophosphates of adduct standards based on their retention times, identified on the bases of their UV spectra, and quantified. In addition, to verify their identities the collected adduct fractions were converted to their ring-opened derivatives by treating with NaBH4 (Fig. 2).

OH

N

Ring-opened Acr-dG

dR

O N

N N H

N

N

Acr-dG

N dR

(B)

UV254nm

28

4.1. Kinetics of formation of cyclic adducts (C)

UV254nm

During the 19-day incubation, using conditions chosen to mimic in vivo lipid peroxidation and to increase the yields of adducts for better detection and quantification, the adducts formed from each fatty acid were identified and quantified at different time intervals. Fig. 3 shows the kinetics of formation of Acr-, Cro-, and HNE-dG adducts from each fatty acid. Both the rates and yields of the formation of Acr adducts from the fatty acids were much greater than those from Cro and HNE. In general, the reaction rates showed a near linear increase for Acr- and Cro-dG for the first 100 h of incubation, followed by a gradual decline of formation, a similar pattern was observed for formation of HNE-dG. The kinetics of the formation of each adduct from PUFAs is dictated by the yield of the enal and its reactivity toward dG. Acr is a major product of both ␻-3 and ␻-6, and is the most reactive one. The reactivity of enals toward dG decreases with the increasing chain length Acr > Cro > Pen > Hep > HNE. It was reported previously that upon oxidation by Fe(II) and H2 O2 , ␻-6 PUFAs yielded comparable amounts of Acr and HNE as major products [5]. However, HNE is much less reactive toward dG than

16

18

20

22

24

26

28

30

32

34

Retention time (min) Fig. 2. Confirmation of identities of Acr-dG adducts by ring-opening/reduction with NaBH4 at pH 11–12. (A) UV standards of Acr-dG 3 and its ring-opened product. (B) AcrdG 3 collected from incubation with AA. (C) Conversion of Acr-dG 3 to a product that co-migrated with the ring-opened product after reaction with NaBH4 at pH 11–12.

Acr, thus, the yields of HNE-dG adducts were considerably less than those of Acr-dG adducts. The low yields of Cro-, Pen-, and Hep-dG adducts are also consistent with the low yields and the relatively low

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

4.2. Newly identified cyclic dG adducts from ω-3 and ω-6 fatty acids

nmol Acr-dG/ mol PUFA

6 DHA EPA LNA AA LA

5 4 3 2 1 0

0

100

200

300

400

500

time, hours

(A)

nmol Cro-dG/ mol PUFA

1.0 DHA EPA LNA

0.8

0.4 0.2

0

100

(B)

200

300

400

500

time, hours

nmol HNE-dG/ mol PUFA

0.4 AA

0.3

0.2

0.1

0.0

(C)

Two previously unidentified 1,N2 -propanodG adducts, Pen- and Hep-dG adducts, were also detected in these reactions. The former are found in the incubation with all three ␻-3 PUFAs, whereas the latter are formed only from AA. The identities of these new 1,N2 -propanodG adducts were confirmed by comparing their UV spectra and retention times with those of the synthetic standards, and further verified by the ring-opening reduction reaction with NaBH4 . The yields of Pen- and Hep-dG were comparable to those of Cro- and HNE-dG adducts from ␻-3 and ␻-6 fatty acids, respectively, and were considerably lower than that of Acr-dG adducts. 4.3. Stereochemical formation of isomeric Acr-dG adduct and the importance of double bonds in PUFAs for Acr-dG adduct formation

0.6

0.0

29

0

100

200

300

400

500

time, hours

Fig. 3. Kinetics of formation of (A) Acr-dG; (B) Cro-dG; and (C) HNE-dG adducts upon incubation with ␻-3 and ␻-6 PUFAs.

reactivity of the corresponding enals. This is further illustrated by the observation that Oct adducts were not detected in the incubation mixture, even though it is one of the major enals formed by oxidation of ␻-6 PUFAs [4].

The double bonds in PUFAs are essential for the formation of cyclic adducts, because no adducts were detected upon incubation with stearic acid under the same conditions. The ␻-3 fatty acids, DHA, EPA, and LNA, are the main sources of Acr-dG adducts. Acr-dG adducts are also formed from AA and LA, but the overall yields are considerably less. The order of formation of Acr adducts from the fatty acids studied is: DHA > EPA > AA > LNA > LA (Fig. 4A). Therefore, the formation of Acr-dG adducts is proportional to the number of double bonds in PUFAs. It is also interesting to note that among the isomers of Acr-dG adducts, only Acr-dG 3 was detected in the reactions with PUFAs. The reason for the preferential formation of Acr-dG 3 from oxidized PUFAs is not known, and is worthy of further investigation. Our early studies showed that reaction of Acr with dG yielded three isomeric adducts (Fig. 1). Adducts 1 and 2 possess a propano ring with the hydroxyl group adjacent to the N2 of guanine (␣-OHPdG), whereas the Acr-dG 3 (␥-OHPdG) is a regio-isomer resulting from ring-closure from the opposite direction. Importantly, Acr-dG 3 was also the predominant adduct detected in vivo using the 32 P-postlabeling/HPLC method [7]. The results of the present study again corroborate the role of PUFAs for the formation of Acr-dG 3 in vivo, and suggest that tissue fatty acids may exert a yet

30

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

Fig. 4. PUFAs and enals as sources for the formation of Acr-, Cro, Pen-, Hep-, and HNE-dG adducts: (A) yields of adducts from ␻-3 and ␻-6 PUFAs; the asterisks indicate that the adducts were not detected and (B) PUFAs as sources for cyclic adducts; the asterisks indicate that the adducts have been detected in vivo.

undefined steric effect toward the formation of Acr-dG adducts. The significance and chemical basis for the observed stereoselectivity needs further investigation. This is especially true in view of the differences in the mutagenic potential of the isomeric adduct of Acr [14]. Unlike Acr adducts, Cro- and HNE-dG adducts are formed by ring-closure in only one direction. Therefore, no stereoselectivity was evident in their formation from oxidized PUFAs. 4.4. ω-3 versus ω-6 fatty acids as sources of enal-derived cyclic adducts In summary, we have demonstrated the formation of cyclic DNA adducts from oxidation of different types of PUFAs. Long-chain Hep- and HNE-dG adducts are derived only from ␻-6 PUFAs, whereas short-chain Acr, Cro, and Pen-dG cyclic adducts are primarily from ␻-3 PUFAs. The sources of PUFAs and enals for the formation of these cyclic adducts and the yields of the cyclic adducts from different fatty acids are shown in Fig. 4B. To date, we have detected Acr-, Cro-, and HNE-dG adducts in vivo as common endogenous background DNA lesions in rodents and humans. However, the presence of Pen and Hep adducts in tissues has yet to be demonstrated.

5. Mutagenesis and carcinogenesis of the cyclic adducts Evidence supports the hypothesis that cyclic adducts of Acr, Cro, and HNE are mutagenic lesions. Acr and Cro are mutagenic in Salmonella tester strains TA-100 and TA-104 [15]. Acr-dG adducts are detected in DNA isolated from these strains under the conditions in which mutation occurred [16]. Although HNE is not mutagenic in these strains, it induces mutations in V79 Chinese hamster cells [17]. We have also detected by immunoassay that Acrand Cro-dG adducts are present in the DNA of Chinese hamster ovary cells incubated with Acr and Cro [18]. Site-specific and other mutagenesis studies have shown that the cyclic propanodG adducts are promutagenic, leading to base substitution and frame-shift mutations [19–24]. A model 1,N2 -propanodG adduct, without the hydroxyl group in the propano ring moiety, induces G to T and G to A mutations. Recent studies showed that Acr-dG incorporated into Escherichia coli DNA lacks miscoding properties and is excised by nucleotide excision repair [25,26]. It is known that factors such as the structure of adduct, sequence context, and host systems can affect the outcome of such studies. These studies, nevertheless, provide evidence

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

showing that some, if not all, of these PUFA-derived cyclic adducts are potentially mutagenic. The mutational potential of Acr-, Cro-, and HNE-dG adducts in mammalian cells has not been fully characterized. However, the mutational spectrum caused by Cro has been reported in the supF gene of the shuttle vector plasmids pMY189 and PZ189 replicated in human cells [27,28]. These studies showed that a majority of mutations are base substitutions consisting of G to T transversions and G to A transitions located at a few hotspots, most notably G–C pairs in 5 -AAGG-3 or 5 -CCTT-3 sequences. Although not proven, the formation of Cro-dG in those sequences is likely the cause of the mutations observed. HNE may induce mutations by forming cyclic 1,N2 -propanodG adducts at the critical sites of target genes that are different from those induced by Acr or Cro. Presently, there is only circumstantial evidence supporting the role of the cyclic adducts in carcinogenesis. Studies in animals have shown that the cyclic adducts are detected at higher than basal levels in the DNA of target tissues after the treatment with Cro, N-nitrosopyrrolidine (NPYR), and cyclophosphamide [6,29]. NPYR and cyclophosphamide are known carcinogens that produce Cro and Acr, respectively, upon metabolism. A 2-year bioassay showed that F344 rats treated with 0.6 mM NPYR in drinking water developed greater than 80% liver tumor incidence (mostly hepatocellular carcinomas); NPYR also caused a twoto three-fold increase above background in Cro-dG adducts in liver DNA of untreated rats [30,31]. These results suggest that Cro-dG may play a role in the NPYR-induced liver tumorigenesis in rats, although NPYR is also a precursor for other alkylated DNA adducts [31]. In the same bioassay, Cro administered at the same dose level induced only a small but significant incidence of liver tumors (30%, mostly nodules). Unfortunately, Cro-dG adducts were not measured in the liver of the treated rats in the study. These results showed that, unlike NPYR, Cro is a weak carcinogen. Furthermore, the low incidence of spontaneous liver tumors in untreated F344 rats also suggests that the background levels of the endogenous cyclic adducts may not be sufficient for tumor development during the lifetime of some animals. HNE did not induce skin tumors in CD-1 mice, whereas, its epoxide was only weakly tumorigenic [32]. Similarly, HNE was inactive as a liver carcinogen in newborn

31

CD-1 mice [32]. In a series of skin tumor bioassays in Sencar mice, the topical application of Acr, Cro, and HNE as initiating agents followed by phorbol ester 12-O-tetradecanoylphorbol-13-acetate promotion failed to induce papillomas, although Cro treatment induced the formation of Cro-dG adducts in the skin DNA [6]. By contrast, we observed a significant increase in skin tumor multiplicity upon initiation with 10 doses (3 ␮g/dose) of benzo[a]pyrene followed by topical treatment of HNE (100 ␮g/dose per 5 days per week for 20 weeks). Taken together, these data support that, while enals may have little or no tumor initiating activity, they seem to act as tumor promoters through their ability to form cyclic DNA adducts. Furthermore, HNE, at concentrations similar to that in tissues (0.1–10 ␮M), stimulated phospholipase C and protein kinase C, both of which prominently control cell division and proliferation [33]. It also modulates gene expression via cell signaling, including activating oncogene c-jun and its transcription factor AP-1 in human hepatic stellate cells [34]. Clearly, additional studies are needed to better define the roles of enals and their cyclic adducts in carcinogenesis. To this end, our recent studies have shed light on the role of HNE-dG adducts in p53 mutation and their repair. The results of these studies are briefly described below. 5.1. Binding and mutation spectrum of HNE in human p53 gene A recent study showed that liver DNA from patients of Wilson’s disease, characterized by hepatic deposition of copper and iron that increases the risk of liver cancer, had a significantly higher mutation frequency in p53 gene hotspots with G to T transversions at codon 249 [35]. The G at the third base of codon 249 represents a mutation hotspot in human cancer, particularly in hepatocellular carcinomas [36]. The same study also demonstrated that exposing a wild-type p53 TK-6 lymphoblastoid cell line to HNE increased predominantly the G to T mutation at codon 249 of the p53 gene, suggesting HNE-dG adducts may play a role in the p53 mutation of Wilson’s disease. Given these results, and the fact that more than 90% of p53 mutations and most of the p53 mutational hotspots in human cancer are located in exons 5, 7, and 8 of the p53 genes, we determined the binding of HNE in this region of p53 gene [37]. DNA

32

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

Fig. 5. HPLC chromatograms obtained from 32 P-postlabeling of the enzymatic hydrolysate of the DNA fragments with HNE-treated sample (A) and control sample (untreated) (B). The upper panel shows the UV standards of HNE-dG and the lower panel is the 32 P-labeled adducts detected in DNA. Because of two chiral carbons in the side chain, two pairs of diastereomers of the ring-opened HNE-dG derivatives are detected by HPLC.

fragment of the p53 gene exon 7 was modified with HNE in a pH7 buffer at 37 ◦ C for 20 h. After purification, the DNA fragment was subjected to enzymatic digestion and analyzed for HNE-dG adducts by the 32 P-postlabeling/HPLC method. The formation of HNE-dG adducts in the DNA fragment of the p53 gene exon 7 was confirmed by the detection of the four isomeric HNE-dG adducts in the HNE-treated DNA, but not in the untreated control (Fig. 5). Subsequent experiments demonstrated that the supercoiled plasmid DNA modified with HNE was cut by UvrABC, whereas the unmodified DNA fragment remained intact when incubated with the nuclease. Fig. 6 shows the gel separation of DNA fragments after UVrABC nuclease incision indicating that HNE–DNA adducts are formed in exon 7 of the p53 gene in a sequence selective manner. The HNE–DNA adducts are preferentially formed at codon 249 (–GAGG∗ C–) of exon 7, and at site X (–CAGG∗ A–) in the intron 7 region (Fig. 6). Similar results, although less striking, were obtained with p53 gene exons 5 and 8 with codons 174 and 286, respectively. The results support the role of HNE as a potential etiological agent and provides evidence that its derived cyclic adducts are mutagenic lesions in the p53 gene in the livers of patients with Wilson’s disease.

5.2. Repair of HNE-dG adducts by human cell nuclear extracts The levels and persistence of the cyclic DNA adducts in tissues are determined by several factors: their rate of formation from enals, the rate of repair of the adducts, and the rate of DNA replication. Studies have shown that cyclic adducts in DNA can be repaired either by base excision (BER) or nucleotide excision repair (NER) mechanisms. The repair of etheno adducts is mainly mediated by the BER pathway, initiated by several different DNA glycosylases, namely N-methylpurine-DNA glycosylase (MPG) and thymine DNA glycosylase present throughout the phylogeny and including humans [38–43]. Other human proteins capable of remo ving etheno adducts have also been characterized [44]. However, 1,N2 -propanodG, a model for bulky cyclic adducts, is not excised by the glycosylases, and instead, is repaired by the NER pathway [45]. It appears that different repair mechanisms are operative for different cyclic adducts. Although, it has been recently shown that Acr-dG adducts are repaired by the NER pathway in E. coli [25,26], no data on the repair of Acr-, Cro-, and HNE-dG adducts in mammalian cells are available. To investigate the recognition and

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

33

Fig. 6. UvrABC incision of HNE-modified DNA fragments (A) and (B); and DNA binding spectrum of HNE in exon 7 of human p53 gene (C). (A) The 5 -32 P-end-labeled 187 bp fragment of human p53 gene exon 7 was modified with (lanes 4 and 5) or without (lane 3) HNE (30 mg/ml), and then reacted with (lanes 3 and 5) or without (lane 4) UvrABC nuclease. (B) The 3 -32 P-end-labeled 141 bp fragment of the p53 gene exon 7 was modified with (lanes 4 and 5) or without (lane 3) HNE (30 mg/ml) and then reacted with (lanes 3 and 5) or without (lane 4) UvrABC nuclease. The codon numbers corresponding to the UvrABC incision bands are depicted on the right side of the panels; X corresponds to a strong binding site in intron 7 of the p53 gene. TC and AG are the Maxam–Gilbert sequencing reactions. (C) DNA binding spectrum in exon 7 of human p53 gene modified with HNE. The 5 -end-labeled (upper panel) or 3 -end-labeled (lower panel) DNA fragments of the p53 gene exon 7 were modified with HNE (30 mg/ml), reacted with UvrABC, and then separated by gel electrophoresis as described in the text. The intensities of UvrABC incision bands in well-separated regions were quantified with a PhosphorImager. The extent of HNE binding is represented by the relative intensity (RI) of UvrABC incision bands. The RI was calculated based on RI = Ij /Imax , where Ij is the intensity of each UvrABC incision band and Imax is the UvrABC incision band with the highest intensity in an autoradiograph.

2000 no cell extract XPA

HNE-dG adduct level (µmol / mol Gua)

repair of HNE-dG adducts by human cell extracts, plasmid DNA substrates containing HNE-dG were prepared by incubating 10 ␮g of pBluescript plasmid DNA with 30 mg/ml HNE at 37 ◦ C for 20 h. The levels of HNE-dG adducts were determined by the 32 P-postlabeling/HPLC method to be ∼1000–1500 adducts/106 dG. This substrate was then used in an in vitro repair-synthesis assay to study the complete repair of HNE-induced DNA adducts by human cell nuclear extracts using the published protocols for repair of UV damage [46]. The levels of HNE-dG in post-repaired DNA substrates were directly measured by the 32 P-postlabeling/HPLC method. The results show that 90% of HNE-dG was repaired by HeLa nuclear extracts in 1 h at 30 ◦ C (Fig. 7). Moreover, under similar repair conditions, HNE-dG was not removed by the nuclear extracts from XPA cells, deficient in NER pathway (Fig. 7). Therefore, HNE-induced

Hela

1000

0 Fig. 7. Levels of HNE-dG in modified plasmid DNA pBlue script II SK− (pBS) after incubation with either no cell extracts, XPA extract, or HeLa extract. The amount of HNE-dG in pBS DNAs after the reaction was analyzed by the 32 P-postlabeling/HPLC method.

34

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

DNA damage, like other cyclic adducts, are repaired in human cells and this repair is accomplished most likely by the NER pathway.

adducts generated from oxidized ␻-6 fatty acids as an etiological agent in human cancers. More evidence from molecular and biological studies is needed in order to better define the role that enal-derived cyclic adducts play in carcinogenesis.

6. Conclusion This study demonstrates that the formation of cyclic DNA adducts from the oxidation of PUFAs depends on the types of fatty acids involved. The short-chain enal-derived cyclic adducts Acr-, Cro-, and Pen-dG are derived primarily from ␻-3 fatty acids, whereas the long chain Hep- and HNE-dG adducts are formed exclusively from ␻-6 fatty acids. Enals are oxidation products of PUFAs [5], and as expected, the enal-derived cyclic adducts were detected only from incubation of fatty acids with multiple double bonds; no adducts were detected from SA, a saturated fatty acid. Consistent with the high reactivity and high yield of Acr, Acr-dG is the predominant product from both ␻-3 and ␻-6 PUFAs, in fact, Acr-dG adducts are the only adducts identified which are formed from both ␻-3 and ␻-6 PUFAs. The similar patterns of formation of Cro- and Pen-dG adducts from DHA, EPA and LNA and Hep- and HNE-dG adducts from AA (Fig. 4) also suggest a common pathway for the formation of Cro and Pen by ␻-3 PUFAs and Hep and HNE by ␻-6 PUFAs. Animal bioassays showed that ␻-6 PUFAs promote tumorigenesis in the colon and mammary glands, whereas ␻-3 PUFAs are protective [47]. However, the mechanisms for their contrasting effects are not yet understood. Distinct patterns of cyclic adduct formation from different PUFAs raises a possibility that the specificity of adduct formation may contribute to the differential effects of these fatty acids. It has been reported that Acr-dG 3 in DNA, due to its facile ring opening, does not cause mispairing of bases, and therefore is a non-mutagenic lesion in vivo [14,25,26,48]. Although the mutational properties of the other cyclic adducts need to be characterized in order to fully assess their potential roles in tumorigenesis, it is interesting to recognize that, in agreement with the published mutational spectrum of HNE in the p53 gene, the HNE-derived cyclic adducts are shown to preferentially bind to the mutation hotspot at the codon 249 in the human p53 gene. These results suggest a possible role of HNE and its cyclic

References [1] F.-L. Chung, R. Young, S.S. Hecht, Formation of cyclic 1,N2 -propanodeoxyguanosine adducts in DNA upon reaction with acrolein or crotonaldehyde, Cancer Res. 44 (1984) 990– 995. [2] R.G. Nath, F.-L. Chung, Detection of exocyclic 1,N2 -propanodeoxyguanosine adducts as common DNA lesions in rodents and humans, Proc. Natl. Acad. Sci. U.S.A. 91 (1994) 7491–7495. [3] R.G. Nath, J.E. Ocando, F.-L. Chung, Detection of 1,N2 -propano-deoxyguanosine adducts as potential endogenous DNA lesions in rodent and human tissues, Cancer Res. 56 (1996) 452–456. [4] F.-L. Chung, R.G. Nath, J.E. Ocando, A. Nishikawa, L. Zhang, Deoxyguanosine adducts of t-4-hydroxy-2-nonenal are endogenous DNA lesions in rodents and humans: detection and potential sources, Cancer Res. 60 (2000) 1507–1511. [5] H.-Y. Wu, J.-K. Lin, Determination of aldehydic lipid peroxidation products with dabsylhydrazine by high-performance liquid chromatography, Anal. Chem. 67 (1995) 1603– 1612. [6] F.-L. Chung, R. Young, S.S. Hecht, Detection of cyclic 1,N2 -propanodeoxyguanosine adducts in DNA of rats treated with N-nitrosopyrrolidine and mice treated with crotonaldehyde, Carcinogenesis 10 (1989) 1291–1297. [7] F.-L. Chung, R.G. Nath, M. Nagao, A. Nishikawa, G.-D. Zhou, K. Randerath, Endogenous formation and significance of 1,N2 -propanodeoxyguanosine adducts, Mutat. Res. 424 (1999) 71–81. [8] J. Nair, H. Stone, M. Nagao, A. Barbin, H. Bartsch, Copper-dependent formation of miscoding etheno-DNA adducts in the liver of Long-Evans Cinnamon (LEC) rats developing hereditary hepatitis and hepatocellular carcinoma, Cancer Res. 56 (1996) 1267–1271. [9] R.G. Nath, J.E. Ocando, J. Richie Jr., F.-L. Chung, Effect of glutathione depletion on exocyclic adduct levels in the liver DNA of F344 rats, Chem. Res. Toxicol. 10 (1997) 1250–1253. [10] J. Nair, C.E. Vaca, I. Velic, M. Mutanen, L.M. Valsta, H. Bartsch, High dietary omega-6 polyunsaturated fatty acids drastically increase the formation of etheno-DNA base adducts in white blood cells of female subjects, Cancer Epid. Biol. Prev. 6 (1997) 597–601. [11] J.-L. Fang, C.E. Vaca, L.M. Valsta, M. Mutanen, Determination of DNA adducts of malonaldehyde in humans: effects of dietary fatty acid composition, Carcinogenesis 17 (1996) 1035–1040. [12] F. El-Ghissassi, A. Barbin, J. Nair, H. Bartsch, Formation of 1,N6 -ethenoadenine and 3,N4 -ethenocytosine by lipid

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

peroxidation products and nucleic acid bases, Chem. Res. Toxicol. 8 (1995) 278–283. J. Pan, F.-L. Chung, Formation of cyclic deoxyguanosine adducts from ␻-3 and ␻-6 polyunsaturated fatty acids under oxidative conditions, Chem. Res. Toxicol. 15 (2002) 367–372. I.-Y. Yang, G. Chan, H. Miller, Y. Huang, M.C. Torres, F. Johnson, M. Moriya, Mutagenesis by acrolein-derived propanodeoxyguanosine adducts in human cells, Biochemistry 41 (2002) 13826–13832. L.J. Marnett, H.K. Hurd, M.C. Hollstein, D.E. Levin, H. Esterbauer, B.N. Ames, Naturally occurring carbonyl compounds are mutagens in Salmonella tester strain TA104, Mutat. Res. 148 (1985) 25–34. P.G. Foiles, S.A. Akerkar, F.-L. Chung, Application of an immunoassay for cyclic acrolein deoxyguanosine adducts to assess their formation in DNA of salmonella typhimurium under conditions of mutation induction by Acr, Carcinogenesis 10 (1989) 87–90. E. Cajelli, A. Ferraris, G. Brambilla, Mutagenicity of 4-hydroxynonenal in V79 Chinese hamster cells, Mutat. Res. 190 (1987) 169–171. P.G. Foiles, S.A. Akerkar, L.M. Miglietta, F.-L. Chung, Formation of cyclic deoxyguanosine adducts in Chinese hamster ovary cells by acrolein and crotonaldehyde, Carcinogenesis 11 (1990) 2059–2061. A.K. Basu, M.L. Wood, L.J. Niedernhofer, L.A. Ramos, J.M. Essigmann, Mutagenic and genotoxic effects of three vinyl chloride-induced DNA lesions: 1,N6 -ethenoAde, 3,N4 -ethenocytosine, and 4-amino-5-(imidazol-2-yl)imidazole, Biochemistry 32 (1993) 12794–12801. M. Moriya, W. Zhang, F. Johnson, A.P. Grollman, Mutagenic potency of exocyclic DNA adducts: marked differences between Escherichia coli and simian kidney cells, Proc. Natl. Acad. Sci. U.S.A. 91 (1994) 11899–11903. G. Pandya, M. Moriya, 1,N6 -Ethenodeoxyadenosine, a DNA adduct highly mutagenic in mammalian cells, Biochemistry 35 (1996) 11487–11492. M. Benamira, U. Singh, L.J. Marnett, Site-specific frameshift mutagenesis by a propanodeoxyguanosine adduct positioned in the (CpG)4 hot-spot of Salmonella typhimurium his D3052 carried on an M13 vector, J. Biol. Chem. 267 (1992) 22392– 22400. K.C. Cheng, B.D. Preston, D.S. Cahill, M.K. Dosanjh, B. Singer, L.A. Loeb, The vinyl chloride DNA derivative N2 ,3-ethenoGua produces G > A transitions in Escherichia coli, Proc. Natl. Acad. Sci. U.S.A. 88 (1991) 9974–9978. V.A. Palejwala, D. Simha, M.Z. Humayun, Mechanisms of mutagenesis by exocyclic DNA adducts, transfection of M13 viral DNA bearing a site-specific adduct shows that ethenocytosine is a highly efficient recA-independent mutagenic noninstructional lesion, Biochemistry 30 (1991) 8736–8743. I.-Y. Yang, M. Hossain, H. Miller, S. Khullar, F. Johnson, A. Grollman, M. Moriya, Responses to the major acrolein-derived deoxyguanosine adduct in Escherichia coli, J. Biol. Chem. 276 (2001) 9071–9076.

35

[26] L.A. Van der Veen, M.F. Hashim, L.V. Nechev, T.M. Harris, C.M. Harris, L.J. Marnett, Evaluation of the mutagenic potential of the principal DNA adduct of acrolein, J. Biol. Chem. 276 (2001) 9066–9070. [27] M. Kawanishi, T. Matsuda, G. Sasaki, T. Yagi, S. Matsu, H. Takebe, A spectrum of mutations induced by crotonaldehyde in shuttle vector plasmids propagated in human cells, Carcinogenesis 19 (1998) 69–72. [28] C. Czerny, E. Eder, T.M. Runger, Genotoxicity and mutagenicity of the ␣,␤-unsaturated carbonyl compound crotonaldehyde on a plasmid shuttle vector, Mutat. Res. 407 (1998) 125–134. [29] V.L. Wilson, P.G. Foiles, F.-L. Chung, A.C. Povey, A.A. Frank, C.C. Harris, Detection of acrolein and crotonaldehyde DNA adducts in cultured human cells and canine peripheral blood lymphocytes by 32 P-postlabeling and nucleotide chromatography, Carcinogenesis 12 (1991) 1483–1490. [30] F.-L. Chung, T. Tanaka, S.S. Hecht, Induction of liver tumors in F344 rats by crotonaldehyde, Cancer Res. 46 (1986) 1285– 1289. [31] F.-L. Chung, M.Y. Wang, S.S. Hecht, Detection of exocyclic Gua adducts in hydrolysates of hepatic DNA of rats treated with N-nitrosopyrrolidine and in calf thymus DNA reacted with ␣-acetoxy-N-nitrosopyrrolidine, Cancer Res. 49 (1989) 2034–2041. [32] F.-L. Chung, H.-J.C. Chen, J.B. Guttenplan, A. Nishikawa, G.C. Hard, 2,3-Epoxy-4-hydroxynonanal as a potential tumor-initiating agent of lipid peroxidation, Carcinogenesis 14 (1993) 2073–2077. [33] H. Esterbauer, R.J. Schaur, H. Zollner, Chemistry and biochemistry of 4-hydroxynonenal, Free Rad. Biol. Med. 11 (1991) 81–128. [34] M.U. Dianzani, 4-Hydroxynonenal and cell signaling, Free Radical Res. 28 (1998) 553–560. [35] S.P. Hussain, K. Raja, P.A. Amstad, M. Sawyer, L.J. Trudel, G.N. Wogan, L.J. Hofseth, P.G. Shields, T.R. Billiar, C. Trautwein, T. Hohler, P.R. Galle, D.H. Phillips, R. Markin, A.J. Marrogi, C.C. Harris, Increased p53 mutation load in nontumorous human liver of Wilson disease and hemochromatosis: oxyradical overload diseases, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 12770–12775. [36] M.S. Greenblatt, W.P. Bennett, M. Hollstein, C.C. Harris, Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis, Cancer Res. 54 (1994) 4855–4878. [37] W. Hu, Z. Feng, J. Eveleigh, G. Iyer, J. Pan, S. Amin, F.-L. Chung, M.-S. Tang, The major lipid peroxidation product, trans-4-hydroxy-2-nonenal, preferentially forms DNA adducts at codon 249 of human p53 gene, a unique mutational hotspot in hepatocellular carcinoma, Carcinogenesis 23 (2002) 1781– 1789. [38] B. Singer, A. Antoccia, A.K. Basu, M.K. Dosanjh, H. Fraenkel-Conrat, P.E. Gallagher, J.T. Kusmierek, Z.-H. Qiu, B. Rydberg, Both purified human 1,N6 -ethenoAde-binding protein and purified human 3-methylAde-DNA glycosylase act on 1,N6 -ethenoAde and 3-methylAde, Proc. Natl. Acad. Sci. U.S.A. 89 (1992) 9386–9390.

36

F.-L. Chung et al. / Mutation Research 531 (2003) 25–36

[39] M.K. Dosanjh, A. Chenna, E. Kim, H. Fraenkel-Conrat, L. Samson, B. Singer, All four known cyclic adducts formed by the vinyl chloride metabolite chloroacetaldehyde are released by a human DNA glycosylase, Proc. Natl. Acad. Sci. U.S.A. 91 (1994) 1024–1028. [40] M.K. Dosanjh, R. Roy, S. Mitra, B. Singer, 1,N6 -Ethenoadenine is preferred over 3-methyladenine as substrate by a cloned human N-methylpurine-DNA glycosylase (3-methyladenine-DNA glycosylase), Biochemistry 33 (1994) 1624– 1628. [41] R. Roy, T. Biswas, J.C. Lee, S. Mitra, Mutation of a unique aspartate residue abolishes the catalytic activity but not substrate binding of the mouse N-methylpurine-DNA glycosylase (MPG), J. Biol. Chem. 275 (2000) 4278–4282. [42] M. Saparbaev, K. Kleibl, J. Laval, Escherichia coli, saccharomyces cerevisiae, rat and human 3-methyladenine DNA glycosylases repair 1,N6 -ethenoadenine when present in DNA, Nucleic Acids Res. 23 (1995) 3750–3755. [43] Z. Matijasevic, M. Sekiguchi, D.B. Ludlum, Release of N2 ,3-ethenoGua from chloroacetaldehyde-treated DNA by

[44]

[45]

[46]

[47]

[48]

Escherichia coli 3-methylAde DNA glycosylase II, Proc. Natl. Acad. Sci. U.S.A. 89 (1992) 9331–9334. B.H. Hang, A. Chenna, S. Rao, B. Singer, 1,N6 -EthenoAde and 3,N4 -ethenocytosine are excised by separate human DNA glycosylases, Carcinogenesis 17 (1996) 155–157. K.A. Johnson, S.P. Fink, L.J. Marnett, Repair of propanodeoxyguanosine by nucleotide excision repair in vivo and in vitro, J. Biol. Chem. 17 (1997) 11434–11438. K. Kohno, T. Shimamoto, Nucleotide excision repair assay in Drosophila melanogaster using established cell lines, DNA repair protocols eukaryotic systems, in: D.S. Henderson (Ed.), Methods in Molecular Biology, vol. 113, Humanae Press, Clifton, UK, 1977, pp. 337–346. C.V. Rao, Y. Hirose, C. Indranie, B.S. Reddy, Modulation of experimental colon tumorigenesis by types and amounts of dietary fatty acids, Cancer Res. 61 (2001) 1927–1933. C. de los Santos, T. Zaliznyak, F. Johnson, NMR characterization of a DNA dup lex containing the major acrolein-derived deoxyguanosine adduct ␥-OH-1,-N2 -propano-2 -deoxyguanosine, J. Biol. Chem. 276 (2001) 9077–9082.