Mutation Research, 250 (1991) 135-144 © 1991 Elsevier Science Publishers B.V. All rights reserved 0027-5107/91/$03.50 ADONIS 0027510791001711
135
MUT 02515
Bulky adducts detected by 32P-postlabeling in DNA modified by oxidative damage in vitro. Comparison with rat lung I-compounds K u r t R a n d e r a t h a, P e i - F a n g Y a n g a, T r a c y F. D a n n a a, R a n j a n i R e d d y a, W i l l i a m P. W a t s o n b a n d E r i k a R a n d e r a t h a a Division of Toxicology, Department of Pharmacology, Baylor College of Medicine, Houston, TX 77030 (U.S.A.) and b Shell Research Ltd., Sittingboume Research Centre, Sittingbourne, Kent ME9 8AG (Great Britain) (Accepted 5 April 1991)
Keywords: Oxidative DNA damage; Oxygen free radicals; DNA adducts; l-compounds; 32p-Postlabelling;Thin-layer chromatography; Fenton reaction
Summary Oxygen free radicals, such as the hydroxyl radical generated by interaction of Fe 2÷ and H 2 0 2 (Fenton reaction), are produced in mammalian cells as a result of aerobic metabolism and under various pathological conditions and are known to elicit mutations and potentially other adverse effects by reacting with DNA bases. Several products thus formed have recently been characterized as hydroxylated derivatives of cytosine, thymine, adenine, and guanine and imidazole-ring-opened derivatives of adenine and guanine in DNA. As shown herein by 3ap-postlabeling, incubation of DNA under Fenton reaction conditions led to additional products which, by virtue of resistance to nuclease P1 catalyzed 3'-dephosphorylation and chromatographic behavior, appeared to be bulky adducts rather than small polar, hydroxylated or ring-opened nucleotide derivatives. Two major and five minor DNA derivatives were measured after 32P-postlabeling and TLC mapping of DNA oxidized in vitro under conditions known to lead to formation of reactive oxygen species. Amounts of products formed depended on Fe 2+ and H 2 0 2 concentrations and increased in the presence of L-ascorbic acid. One of the two major products was also detected in lung DNA of rats where its amount increased with animal age. Thus, at least one l-compound appeared to have its origin in the interaction of DNA with reactive oxygen species.
During recent years, there has been increasing interest in the damage by reactive oxygen species to biological systems and the underlying molecu-
Correspondence: Dr. K. Randerath, Division of Toxicology, Department of Pharmacology, Baylor College of Medicine, Houston, TX 77030 (U.S.A.).
lar mechanisms (reviewed in Halliwell and Gutteridge (1985), Sies (1985), Fridovich (1986), and Halliwell and Gutteridge (1986)). Reactive oxygen species, including hydroxyl radical ( O H ) , hydrogen peroxide ( H 2 0 2 ) , and superoxide anion ( 0 2 • -), are generated in vivo as by-products of normal aerobic metabolism and by phagocytic cells in response to bacterial invasion, as well as
136
by ionizing radiation, ischemia, and "rcdox" cycling agents such as paraquat and certain antineoplastic drugs. While the organism is protected against oxygen radical damage by antioxidant defense systems, these systems may be ovcrwhelmed under conditions of oxidative stress, resulting in injury. Various processes (Ames, 1983; Halliwell, 1987: Floyd, 19901, including degenerative and inflammatory disorders, tissue damage by ischemia/reperfusion, cancer, and aging, are now considered to involve reactive oxygen species. Experimental evidence implies a role of oxygen frec radicals in initiation (reviewed by Marnctt (1987). Floyd (1990) and Breimer (1990)) and promotion of carcinogenesis (Slaga et al., 1981; Kensler and Trush, 1984; Cerutti, 1985). Initiating factors include ionizing radiation (Kennedy et al., 1984), foreign bodies such as asbestos (Mossman et al., 1990), transition metals (Li et al., 1987; Kasprzak and Hernandez, 1989: Kawanishi ct al., 1989; Umemura et al., 1990; Zhong et al., 19901, certain indirectly acting carcinogens such as peroxisome proliferators (Kasai et al., 1989), and choline devoid diet (Rushmore et al., 1987). The involvement of oxygen free radicals in tumor initiation implies their interaction with genomic DNA. In studies of effects of oxygen free radicals on DNA in vitro, thc Fenton reaction, in which H 2 0 2 is reduced by Fe -~+ to hydroxyl radical, or modifications thereof have been used frequently (Aruoma et al., 1989a, b; Floyd, 1990). Such systems, ionizing radiation, or h y p o x a n t h i n e / xanthine oxidase combined with iron ions give rise to oxygen free radicals which produce many different lesions in DNA, such as base modifications, base-free sites, and strand breaks (Breimer, 1990). Three types of base modifications generated by the above systems have been characterized, i.e. (i) the hydroxylation products thymine glycol and cytosine glycol containing a saturated ring, (ii) the hydroxylation products 8-hydroxyguanine, 8-hydroxyadenine, and 5,6-dihydroxycytosinc with intact aromatic ring systems, and (iii) the imidazole-ring-opened products 2,6-diamino4-hydroxy-5-formamidopyrimidine and 4,6-diamino-5-formamidopyrimidine (Aruoma et al.. 1989a, b). Levels of individual lesions depend on the oxygen radical generating system. Mutations were found to occur in Fe e. ex-
posed phage X174 am3 DNA when transfcctcd into E. coil sphcroplasts (Loeb et all., 1988). Various types of oxygen free radical mediated mutations, including basc changes, deletions, and frameshifts, have been demonstrated at the .supF gene in a shuttle vector (Moracs el al., 1989, 199(I) and the aprt locus (Miles and Meuth, 19891 following He()~ exposure or -,/-irradiation. Gcnc rearrangements have bccn observed in a number of mammalian systems upon exposure to ionizing radiation (reviewed in Breimer (1988)and Miles and Mouth (1989)). One of the most studied mutagenic base lesions induced in vivo by reactive oxygen species is 8-hydroxyguaninc (8-Ot I-G) (Kasai and Nishimura, 1984: Kasai ct al., 1986: Floyd, 199(I). This base modification has been shown to miscodc dircctly in model cxpcriments and induce miscoding at neighboring bases (Kuchino ctal., 19871. Associations betwecn carcinogenesis mediated by reactive oxygen spccics and 8-OH-G formation in target tissue I)NA have been reported for the hepatocarcinogcnic, non-mutagenic pcroxisomc prolifcrator ciprofibrate (Kasai et al., 1989), betel quid (Nair et al.. 1990), and thc kidney carcinogens potassium bromate (Kasai ct al., 1987), ferric nitrilotriacetatc (Umemura el al., 1990). and nickel salts (Kasprzak ctal., 1990). 8-OH-G is reported to accumulate, also, in rat organ DNA during aging (Fraga ct al.. 1990). While involvement of reactive oxygen species in carcinogenesis and other disease processes has been established, further work is needed to elucidate the underlying mechanisms by defining the critical intermediates and macromolect, lar lesions. In this context, wc have addressed the question as to whether reactive oxygen species, in addition to inducing small basc modifications, also give rise to bulky DNA adducts. As shown herein, such adducts were indeed detected by -P-post[abehng (Randerath ct al.. 1981, 1t,~85: Gupta ctal., 1982; Reddy et al., 1984: Reddy and Randerath, 1986) in DNA exposed in vitro to Fcnton reaction conditions (Fc'" and H , O , ) . In addition, one of the major spots formed in vitro was also detected in rat lung DNA in viw~ where it increased with agc, thus fitting thc definition of an I-compound (Randcrath ct al.. 1986, 1988, 1989, 199111.
137
Materials and methods
TABLE 1
Materials
SOLVENTS FOR ADDUCT SEPARATION BY PE1-CELLULOSE TLC
H 2 O 2 (certified ACS) was purchased as a 30% solution from Fisher Scientific (Pittsburgh, PA). Iron(II) sulfate (ACS) was from Alfa Products (Danvers, MA). EDTA, assayed as 99%, and L-ascorbic acid (ACS), assayed as 99.5%, were obtained from Sigma Chemical Co. (St. Louis, MO). Female Sprague-Dawley rats of different ages were from Harlan Sprague-Dawley Inc. (Houston, TX). Lung DNA was isolated from 1, 4 and 9 month old female Sprague-Dawley rats (two or three per group) by a procedure involving solvent extraction and enzymatic digestion of protein and RNA (Gupta, 1984). All reagent solutions for DNA modification were freshly prepared. The sources of materials for ~2p-postlabeling analysis have been previously reported (Randerath et al., 1981; Gupta et al., 1982; Reddy et al., 1984; Reddy and Randerath, 1986). PEIcellulose TLC sheets were prepared in the laboratory (Randerath and Randerath, 1966).
~2P-Postlabeling analysis of DNA Conditions for 32p-postlabeling were similar as previously described (Randerath et al., 1981; Reddy et al., 1981; Gupta et al., 1982). 10/xg of DNA was hydrolyzed to nucleoside 3'-monophosphates with a mixture of micrococcal nucleasc (40 mU//zl) and spleen phosphodiesterase (0.4 /zg/ tzl). The adducts were enriched by nuclease P1 (0.6 g g / g l ) treatment, 32p-labeled, and mapped by TLC as 3',5'-bisphosphates (Reddy and Randerath, 1986). Adducts were first purified by PEt-cellulose TLC with solvent I (Table l). Adducts retained at and slightly above the origin (lower cuts as defined by Randerath et al. (1988)) were then contact-transferred to a fresh PEl-celluiose sheet and separated by two-dimensional TLC with solvent It, and finally solvent IIl. In some experiments reductions (up to 15%) of solvent concentrations led to improved recoveries of fast moving nucleotides. Autoradiographic detection was conducted at -80°C for 3 h to 3 days, depending on the amount of radioactivity, and employed Kodak XAR-5 film and DuPont Lightning Plus intensifying screens. Control and modified samples to be compared in the same experi-
Solvent
Composition
I
2.3 M sodium phosphate, pH 5.7
II
3.31 M lithium formate, 5.9 M urea, ptt 3.4
III "
(I.56 M sodium phosphate, 5.6 M urea, pH 6.4
IV
0.55 M NaCI, 0.23 M sodium phosphate, 0.39 M Tris, 6.7 M urea, pl| 4.5
g
a
0.21 M sodium phosphate, 2.1 M urea, pH 6.4
VI
0.28 M ammonium sulfate, 50 mM sodium phosphate, pH 6.8
VII
0.38 M Tris-HCI, 0.38 M H3BO 3, 7.5 mM EDTA, 0.98 M NaCI, 6.0 M urea, pH 8.0
Vlll
(I.28 M sodium phosphate, 0.18 M Tris, 3.0 M urea, pH 8.2
IX
isopropanol/4 N ammonium hydroxide (50:50,
v/v) " Predeveloped to 2 cm with urea-free phosphate of the same strength.
ment were autoradiographed under identical conditions.
Production of bulky DNA adducts by Fenton-type reaction and comparison with lung DNA I-compounds Lung DNA from 1, 4 and 9 month old female Sprague-Dawley rats was assayed in triplicate for l-compounds by nuclease P1 enhanced 32p-postlabeling (Reddy and Randerath, 1986). To ensure comparability of results from the three age groups, tissues and DNA were isolated and analyses performed in parallel. Rat-lung DNA (0.1 g g / ~ l ) was incubated similarly as described by Lesko et al. (1982) employing 600 #M Fe 2÷, 600 ~M EDTA, pH 5.5, and 3 mM H20 2 at 370C for 1 h. DNA was precipitated at 0 ° C by the addition of 0.1 vol. of 5 M NaCI and 1.1 voi. of absolute ethanol, then washed twice with 70% ethanol. The recovery was determined spectrophotometrically assuming 1 mg = 20 A2~~ units. The chemically induced spot pattern was compared with lung DNA I-compounds by .42P-postlabeling and co-chromatography. To co-chromatograph lung I-spot c (Fig. 1)
138
c
135
MUT 02515
Bulky adducts detected by 32P-postlabeling in DNA modified by oxidative damage in vitro. Comparison with rat lung I-compounds K u r t R a n d e r a t h a, P e i - F a n g Y a n g a, T r a c y F. D a n n a a, R a n j a n i R e d d y a, W i l l i a m P. W a t s o n b a n d E r i k a R a n d e r a t h a a Division of Toxicology, Department of Pharmacology, Baylor College of Medicine, Houston, TX 77030 (U.S.A.) and b Shell Research Ltd., Sittingboume Research Centre, Sittingbourne, Kent ME9 8AG (Great Britain) (Accepted 5 April 1991)
Keywords: Oxidative DNA damage; Oxygen free radicals; DNA adducts; l-compounds; 32p-Postlabelling;Thin-layer chromatography; Fenton reaction
Summary Oxygen free radicals, such as the hydroxyl radical generated by interaction of Fe 2÷ and H 2 0 2 (Fenton reaction), are produced in mammalian cells as a result of aerobic metabolism and under various pathological conditions and are known to elicit mutations and potentially other adverse effects by reacting with DNA bases. Several products thus formed have recently been characterized as hydroxylated derivatives of cytosine, thymine, adenine, and guanine and imidazole-ring-opened derivatives of adenine and guanine in DNA. As shown herein by 3ap-postlabeling, incubation of DNA under Fenton reaction conditions led to additional products which, by virtue of resistance to nuclease P1 catalyzed 3'-dephosphorylation and chromatographic behavior, appeared to be bulky adducts rather than small polar, hydroxylated or ring-opened nucleotide derivatives. Two major and five minor DNA derivatives were measured after 32P-postlabeling and TLC mapping of DNA oxidized in vitro under conditions known to lead to formation of reactive oxygen species. Amounts of products formed depended on Fe 2+ and H 2 0 2 concentrations and increased in the presence of L-ascorbic acid. One of the two major products was also detected in lung DNA of rats where its amount increased with animal age. Thus, at least one l-compound appeared to have its origin in the interaction of DNA with reactive oxygen species.
During recent years, there has been increasing interest in the damage by reactive oxygen species to biological systems and the underlying molecu-
Correspondence: Dr. K. Randerath, Division of Toxicology, Department of Pharmacology, Baylor College of Medicine, Houston, TX 77030 (U.S.A.).
lar mechanisms (reviewed in Halliwell and Gutteridge (1985), Sies (1985), Fridovich (1986), and Halliwell and Gutteridge (1986)). Reactive oxygen species, including hydroxyl radical ( O H ) , hydrogen peroxide ( H 2 0 2 ) , and superoxide anion ( 0 2 • -), are generated in vivo as by-products of normal aerobic metabolism and by phagocytic cells in response to bacterial invasion, as well as
136
by ionizing radiation, ischemia, and "rcdox" cycling agents such as paraquat and certain antineoplastic drugs. While the organism is protected against oxygen radical damage by antioxidant defense systems, these systems may be ovcrwhelmed under conditions of oxidative stress, resulting in injury. Various processes (Ames, 1983; Halliwell, 1987: Floyd, 19901, including degenerative and inflammatory disorders, tissue damage by ischemia/reperfusion, cancer, and aging, are now considered to involve reactive oxygen species. Experimental evidence implies a role of oxygen frec radicals in initiation (reviewed by Marnctt (1987). Floyd (1990) and Breimer (1990)) and promotion of carcinogenesis (Slaga et al., 1981; Kensler and Trush, 1984; Cerutti, 1985). Initiating factors include ionizing radiation (Kennedy et al., 1984), foreign bodies such as asbestos (Mossman et al., 1990), transition metals (Li et al., 1987; Kasprzak and Hernandez, 1989: Kawanishi ct al., 1989; Umemura et al., 1990; Zhong et al., 19901, certain indirectly acting carcinogens such as peroxisome proliferators (Kasai et al., 1989), and choline devoid diet (Rushmore et al., 1987). The involvement of oxygen free radicals in tumor initiation implies their interaction with genomic DNA. In studies of effects of oxygen free radicals on DNA in vitro, thc Fenton reaction, in which H 2 0 2 is reduced by Fe -~+ to hydroxyl radical, or modifications thereof have been used frequently (Aruoma et al., 1989a, b; Floyd, 1990). Such systems, ionizing radiation, or h y p o x a n t h i n e / xanthine oxidase combined with iron ions give rise to oxygen free radicals which produce many different lesions in DNA, such as base modifications, base-free sites, and strand breaks (Breimer, 1990). Three types of base modifications generated by the above systems have been characterized, i.e. (i) the hydroxylation products thymine glycol and cytosine glycol containing a saturated ring, (ii) the hydroxylation products 8-hydroxyguanine, 8-hydroxyadenine, and 5,6-dihydroxycytosinc with intact aromatic ring systems, and (iii) the imidazole-ring-opened products 2,6-diamino4-hydroxy-5-formamidopyrimidine and 4,6-diamino-5-formamidopyrimidine (Aruoma et al.. 1989a, b). Levels of individual lesions depend on the oxygen radical generating system. Mutations were found to occur in Fe e. ex-
posed phage X174 am3 DNA when transfcctcd into E. coil sphcroplasts (Loeb et all., 1988). Various types of oxygen free radical mediated mutations, including basc changes, deletions, and frameshifts, have been demonstrated at the .supF gene in a shuttle vector (Moracs el al., 1989, 199(I) and the aprt locus (Miles and Meuth, 19891 following He()~ exposure or -,/-irradiation. Gcnc rearrangements have bccn observed in a number of mammalian systems upon exposure to ionizing radiation (reviewed in Breimer (1988)and Miles and Mouth (1989)). One of the most studied mutagenic base lesions induced in vivo by reactive oxygen species is 8-hydroxyguaninc (8-Ot I-G) (Kasai and Nishimura, 1984: Kasai ct al., 1986: Floyd, 199(I). This base modification has been shown to miscodc dircctly in model cxpcriments and induce miscoding at neighboring bases (Kuchino ctal., 19871. Associations betwecn carcinogenesis mediated by reactive oxygen spccics and 8-OH-G formation in target tissue I)NA have been reported for the hepatocarcinogcnic, non-mutagenic pcroxisomc prolifcrator ciprofibrate (Kasai et al., 1989), betel quid (Nair et al.. 1990), and thc kidney carcinogens potassium bromate (Kasai ct al., 1987), ferric nitrilotriacetatc (Umemura el al., 1990). and nickel salts (Kasprzak ctal., 1990). 8-OH-G is reported to accumulate, also, in rat organ DNA during aging (Fraga ct al.. 1990). While involvement of reactive oxygen species in carcinogenesis and other disease processes has been established, further work is needed to elucidate the underlying mechanisms by defining the critical intermediates and macromolect, lar lesions. In this context, wc have addressed the question as to whether reactive oxygen species, in addition to inducing small basc modifications, also give rise to bulky DNA adducts. As shown herein, such adducts were indeed detected by -P-post[abehng (Randerath ct al.. 1981, 1t,~85: Gupta ctal., 1982; Reddy et al., 1984: Reddy and Randerath, 1986) in DNA exposed in vitro to Fcnton reaction conditions (Fc'" and H , O , ) . In addition, one of the major spots formed in vitro was also detected in rat lung DNA in viw~ where it increased with agc, thus fitting thc definition of an I-compound (Randcrath ct al.. 1986, 1988, 1989, 199111.
137
Materials and methods
TABLE 1
Materials
SOLVENTS FOR ADDUCT SEPARATION BY PE1-CELLULOSE TLC
H 2 O 2 (certified ACS) was purchased as a 30% solution from Fisher Scientific (Pittsburgh, PA). Iron(II) sulfate (ACS) was from Alfa Products (Danvers, MA). EDTA, assayed as 99%, and L-ascorbic acid (ACS), assayed as 99.5%, were obtained from Sigma Chemical Co. (St. Louis, MO). Female Sprague-Dawley rats of different ages were from Harlan Sprague-Dawley Inc. (Houston, TX). Lung DNA was isolated from 1, 4 and 9 month old female Sprague-Dawley rats (two or three per group) by a procedure involving solvent extraction and enzymatic digestion of protein and RNA (Gupta, 1984). All reagent solutions for DNA modification were freshly prepared. The sources of materials for ~2p-postlabeling analysis have been previously reported (Randerath et al., 1981; Gupta et al., 1982; Reddy et al., 1984; Reddy and Randerath, 1986). PEIcellulose TLC sheets were prepared in the laboratory (Randerath and Randerath, 1966).
~2P-Postlabeling analysis of DNA Conditions for 32p-postlabeling were similar as previously described (Randerath et al., 1981; Reddy et al., 1981; Gupta et al., 1982). 10/xg of DNA was hydrolyzed to nucleoside 3'-monophosphates with a mixture of micrococcal nucleasc (40 mU//zl) and spleen phosphodiesterase (0.4 /zg/ tzl). The adducts were enriched by nuclease P1 (0.6 g g / g l ) treatment, 32p-labeled, and mapped by TLC as 3',5'-bisphosphates (Reddy and Randerath, 1986). Adducts were first purified by PEt-cellulose TLC with solvent I (Table l). Adducts retained at and slightly above the origin (lower cuts as defined by Randerath et al. (1988)) were then contact-transferred to a fresh PEl-celluiose sheet and separated by two-dimensional TLC with solvent It, and finally solvent IIl. In some experiments reductions (up to 15%) of solvent concentrations led to improved recoveries of fast moving nucleotides. Autoradiographic detection was conducted at -80°C for 3 h to 3 days, depending on the amount of radioactivity, and employed Kodak XAR-5 film and DuPont Lightning Plus intensifying screens. Control and modified samples to be compared in the same experi-
Solvent
Composition
I
2.3 M sodium phosphate, pH 5.7
II
3.31 M lithium formate, 5.9 M urea, ptt 3.4
III "
(I.56 M sodium phosphate, 5.6 M urea, pH 6.4
IV
0.55 M NaCI, 0.23 M sodium phosphate, 0.39 M Tris, 6.7 M urea, pl| 4.5
g
a
0.21 M sodium phosphate, 2.1 M urea, pH 6.4
VI
0.28 M ammonium sulfate, 50 mM sodium phosphate, pH 6.8
VII
0.38 M Tris-HCI, 0.38 M H3BO 3, 7.5 mM EDTA, 0.98 M NaCI, 6.0 M urea, pH 8.0
Vlll
(I.28 M sodium phosphate, 0.18 M Tris, 3.0 M urea, pH 8.2
IX
isopropanol/4 N ammonium hydroxide (50:50,
v/v) " Predeveloped to 2 cm with urea-free phosphate of the same strength.
ment were autoradiographed under identical conditions.
Production of bulky DNA adducts by Fenton-type reaction and comparison with lung DNA I-compounds Lung DNA from 1, 4 and 9 month old female Sprague-Dawley rats was assayed in triplicate for l-compounds by nuclease P1 enhanced 32p-postlabeling (Reddy and Randerath, 1986). To ensure comparability of results from the three age groups, tissues and DNA were isolated and analyses performed in parallel. Rat-lung DNA (0.1 g g / ~ l ) was incubated similarly as described by Lesko et al. (1982) employing 600 #M Fe 2÷, 600 ~M EDTA, pH 5.5, and 3 mM H20 2 at 370C for 1 h. DNA was precipitated at 0 ° C by the addition of 0.1 vol. of 5 M NaCI and 1.1 voi. of absolute ethanol, then washed twice with 70% ethanol. The recovery was determined spectrophotometrically assuming 1 mg = 20 A2~~ units. The chemically induced spot pattern was compared with lung DNA I-compounds by .42P-postlabeling and co-chromatography. To co-chromatograph lung I-spot c (Fig. 1)
138
c
