Differences in D N A damage induced by mutagenic and nonmutagenic 4-aminoazobenzene derivatives in Escherichia cell
Misaki Kojima a, Toshiteru Morita b, Masakuni Degawa c, Yoshiyuki Hashimoto ¢ and Mariko Tada a
a Laboratory of Biochemistry, Aichi Cancer Center Research Institute, Nagoya 464, b Department of Biology, College of General Education, Osaka University, Osaka 560 and c Department of Hygienic Chemistry, Pharmaceutical Institute, Tohoku Unit'ersity,
Sendal 980 (Japan)
(Received 24 January 1991) (Revision received 27 September 1991)
(Accepted 3 October 1991)
Keywords: N-OH-3-MeO-AAB; N-OH-2-MeO-AAB; Cytotoxic and mutagenic effects; SOS induction; DNA damage
Summary
DNA lesions produced in Escherichia coil AB2500 (uvrA) exposed to the carcinogen N-hydroxy-3- methoxy-4-aminoazobenzene (N-OH-3-MeO-AAB) or the noncarcinogen N-hydroxy-2-methoxy-4- aminoazobenzene (N.OH-2.MeO-AAB) were investigated by alkaline sucrose gradient sedimentation and ~2P-postlabeling analysis. Alkali-labile sites appeared to be formed equally in cells treated with both aminoazobenzene derivatives. 32P-Postlabeling analysis revealed that the 3-MeO-AAB-DNA adduct level was 25-fold higher than that for 2.MeO-AAB-DNA adducts. In addition to major adducts, 4 minor spots were detected in N.OH-3-MeO-AAB-treated cells, while only one major adduct was found in N.OH-2- MeO-AAB-treated cells. The mutagenicities and cytotoxicities were also determined with E. coil with different repair capacities; we found that repair of 3-MeO-AAB damages is strongly dependent on the UVR repair system. Moreover, N-OH-3-MeO-AAB, but not N.OH-2-MeO-AAB, could induce recA and umuC gene expression, which was higher in uvrA strains than in the wild type.
It is known that 3-methoxy-4-aminoazob'en- zone (3.MeO-AAB) is a potent hepatocarcinogen to rats and a potent mutagen to bacteria, whereas 2-methoxy.4.aminoazobenzene (2-MeO-AAB), which differs only in the methoxyi-substituted po- sition of the same benzene ring, is a noncarcino- gen and a nonmutagen (Hashimoto et al., 1981a;
Correspondence: M. Tada, Laboratory of Biochemistry, Aichi Cancer Center Research Institute, Nagoya 464 (Japan).
Miller and Miller, 1961; Odashima and Hashi- mote, 1968; Odashima and Hashimoto, 1970). Our previous reports revealed that (1) 3-MeO- AAB was far more mutagenic than 2-MeO-AAB in the Salmonella mutation system (Hashimoto et al., 1981a), (2) 3-MeO-AAB induced unscheduled DNA synthesis in rat and mouse hepatocytes but 2-MeO-AAB did not (Watanabe and Hashimoto, 1981) and (3) both N-OH-2-MeO-AAB and N- OH-3-MeO-AAB could bind to DNA in vitro by yeast seryl-tRNA synthetase with the same effi-
66
ciency (Hashimoto et ai., 1981b). These findings raised an interesting question as to the com- pletely different nature of DNA lesions produced by structurally similar aminoazo-dye derivatives.
We studied the cytotoxic and mutagenic ef- fects of 2-MeO-AAB and 3-MeO-AAB on E. coli strains with different repair capacities and DNA damages caused by these compounds were ana- lyzed by the alkaline sucrose gradient sedimenta- tion and 32P-postlabeling methods. From our ex- periments, we concluded that the differences in biological potency of these chemicals are due to quantitative and qualitative differences in adduct formation in E. coli.
Materials and methods
Materials N-OH-3-MeO-AAB and N-OH-2-MeO-AAB
(Fig. 1) were synthesized in our laboratory (Hashimoto et al., 1981a). Spleen phosphodi- esterase and micrococcal endonuclease were from Sigma and Boehringer Mannheim, respectively. T4 polynucleotide kinase was obtained from Pharmacia. Polyethyleneimine (PEl)-cellulose TLC sheets were from Schleicher and Schuell. [y-~'P]Adenosine-5'-triphosphate (7000 Ci / mmole) was from ICN gadiochemicals, Irvine, CA. [Methyl-all]thymine (55 Ci/mmole) and [methyl-14C]thymidine (56 Ci/mmole) were ob- tained from Amcrsham Radiochemicals. E. coil K-12 strains, ABlI57 (wildotype), AB2500 (uvrA6), AB2463 (fecAl3) (Bachmann, 1987) and F.. coil B strains H/r30R (wild-type), Hs30R (uvrA), NO30 (recA) (Kondo et al., 1970) were used for the assay of cytotoxicity and mutagenic- ity. For measurement of gene expression, E. coil GE94 (recA.lacZ')and its uvrA6 mutant KY946, and KY700 (umuDC'.lacZ') and its uvrA6 mu- tant KY706 (Yamamoto et al., 1985) were used.
Fig. 1. Chemical structures of N-OH-3-MeO-AAB (left) and N-OH-2-MeO-AAB (right),
Media The rich medium employed was Luria broth
(LB). Nutrient broth consisted of 0.8% Difco nutrient broth and 0.4% NaCl. Semi-enriched medium (SEM) was medium E (Vogel and Bon- ner, 1956) supplemented with 0.4% glucose, 0.04% nutrient broth and 0.02% NaCL To these media, 30 /~g/ml ampicillin was added as re- quired.
Assay for cytotoxicity and mutagenicity E. coli B strains were grown in nutrient broth
at 37°C overnight. The cells were collected and resuspended in M9 buffer to ODt,0o ffi 0.5-0.6 (5 x l0 s cells/ml) containing various concentra- tions of N-OH-2-MeO-AAB or N-OH-3-MeO- AAB dissolved in 20/~l DMSO. The mixture was incubated for 30 rain at 37°C with shaking. To count the surviving cells, reaction mixtures were appropriately diluted with M9 buffer and layered onto SEM agar plates with 0.8% soft agar con- taining 0.7% NaCI. Plates were incubated at 37°C for 1 day and the colonies formed were counted. For measurement of Arg---* Arg + revertant colonies, the cells in the residual reaction mixture were collected and layered on SEM plates with soft agar. After incubation at 37°C for 2 days, the numbers of revertant colonies formed were counted and the mutation frequency was calcu- lated (Kondo et al., 1970).
13.Galactosidase assay ~-Oalactosidase activity was determined by the
method of Miller (1972). The absorbance at 420 nm and 550 nm was measured with a Shimadzu UV 210A Spectrometer. The unit of enzyme ac- tivity was calculated according to the method of Miller (1972).
Alkalb~¢ ~uctose gradient sedimentation The thymine-requiring strain E. coli AB2500
was uniformly labeled with [3H]thymine overnight at YPC (1 /~Ci/ml) in M9 supplemented with 0.25% ¢asamino acids, 1% glucose and 5 /~M thymine. After starvation for 30 rain, the cells (1-5 × 108/mi) were treated with N-OH-3-MeO- AAB or N-OH-2-MeO-AAB in M9 buffer at 37"C for 30 min with shaking. The spheroplasts pre- pared from 0.5 ml of cell suspensions (Ganesan et
al., 1981) were loaded on 100/~! of 0.5 N NaOH layered on the top of 4.8 ml of a 5-20% (w/v) linear sucrose gradient containing 0.1 N NaOH. The loaded gradient was allowed to stand for 60 rain at room temperature, then centrifuged in a Hitachi RST 40T-2 rotor for 90 rain at 28,000 rpm, 20°C. After the run, fractions were collected and counted for acid-precipitable radioactivity on Whatman glass filter (GF/C) in a liquid scintilla- tor. Molecular weights of DNA were calculated as described by Abelson and Thomas (1966) using the value of 15 × 106 Da for a single-stranded DNA of A phage.
3ZP-Postlabeling assay E. coli AB2500 was grown in LB at 37°C,
overnight, collected by centrifugation and resus- pended with M9 buffer at 5 × 10 z° cells in 1 ml. After starvation at 37°C for 30 rain, cells were treated with N-OH-3-MeO-AAB or N-OH-2- MeO-AAB at 37°C for 30 rain. Following incuba- tion, the cells were collected by centrifugation and washed once with 0.15 M NaCI-0.1 M EDTA (pH 8.0). DNA was isolated from the cells by the method of Marmur (1961). DNA (2 pg) was digested by micrococcal endonuclease and spleen phosphodiesterase, and adducts were isolated by extraction with 1-butanol and 5'-32P-labeled with carrier-free [7-32P]ATP by T4 polynucleotide ki- nase according to the method of Gupta (1985).
The 32P-labeled adduct nucleotides were ana- lyzed by multidirectional PEl-cellulose TLC ac- cording to our previously published procedure (Kojima et al., 1991). Adducts were detected by screen-enhanced autoradiographic exposure at - 80oc.
The total nucleotides were analyzed by one-di- rectional PEl-cellulose TLC with 0.27 M ammo- nium sulfate. After detection by autoradiography, adduct and total nucleotide spots were excised from the chromatogram and the radioactivities measured by Cerenkov assay. The adduct levels were calculated by the method of Gupta (1985).
R e s u l t s
Cytotoxic and mutagenic effects o f N-OH-3-MeO- A A B and N-OH-2-MeO-AAB on E. coli
The cytotoxic effects of N.OH-3-MeO-AAB or N-OH-2-MeO-AAB were studied using E. coli B
67
.p
,g
1 0 0
1 0 10
| I i i i i I 12.5 25 50 1 25 50 100
Concentrat ion o f azo dye (UH)
Fig. 2. Cytotoxicity of N-OH-3-MeO-AAB ( ) and N- OH-2-MeO-AAB (--- - - - ) to E. coli B (left) and E. coil
K-12 (right) strains, o. wild; e, ut'rA; z~, recA.
I
. /
t 10 . 7
10 . 8 6.25 12.5 25 50 100
Azo dye concentration (uM)
Fig. 3. Mutation frequency in E. coil B strains with N-OH-3- MeO-AAB ( ) and N-OH-2-MeO-AAB (------). o,
H/r30R; e, Hs30R.
68
and K-12 strains with different DNA repair ca- pacities (Fig. 2). In both E. coli strains, uvrA mutant cells (Hs30R and AB2500) were more sensitive to N-OH-3-MeO-AAB than wild ones (H/r30R and ABl157). On the other hand, the recA mutant cell (NG30 and AB2463) had a sen- sitivity similar to that of the wild-type cells. In contrast to N-OH-3-MeO-AAB, N-OH-2-MeO- AAB exerted only slight killing effects on the strains used in this experiment.
The frequencies of Are + mutation induced by N-OH-3-MeO-AAB or N-OH-2-MeO-AAB in E. coli B strains were also determined. As shown in Fig. 3, the mutation frequencies of N-OH-3- MeO-AAB in H/ r30R (wild-type) and Hs30R (uvrA) increased in a dose-dependent fashion, the value in Hs30R being about 100-fold that in H/r30R. No induced revertants were found in NG30 (recA). On the other hand, revertants in-
duced by N-OH-2-MeO-AAB were only found in Hs30R in a dose-dependent fashion, but only at very low levels by comparison with N-OH-3- MeO-AAB (Fig. 3).
Induction o f recA or umuC gene expression in E. coli by N-OH-3-MeO-AAB or N-OH-2-MeO-AAB
Induction of SOS responses after treatment with N-OH-3-MeO-AAB or N-OH-2-MeO-AAB was determined by measuring the/]-galactosidase activity produced under the transcriptional con- trol of the promoter of the recA gene or the umuC operon in the tester cells. The dose-re- sponse curves of fl-galactosidase activity induced by N-OH-3-MeO-AAB and N-OH-2-MeO-AAB are shown in Fig. 4. Induction of/3-galactosidase activity by N-OH-3-MeO-AAB in uvrA strains (KY946 and KY706) was higher than in wild-type strains (GE94 and KY700). On the other hand,
1 2
';: 8
A j .
i o ~ o " •
, o , , ~ - : - - . . . . . o . . . . . . . . . . . . . . o
0 25 50 100
3 . /
/ 1
" . . . . . . . . . . . . . . . .
0 25 50 100
Concentration o f azo dye (UM) Fig. 4. Induction of recA gene (A) and amuC 8ene (B) expressioa in E, coli by azo dye, Treatment with N-OH-3-MeO-AAB ( ) or N-OH-2-MeO.AAB (--- - - - ) was performed for 90 rain at YPC, o, GE94; o, KY946 (m.rA); A, KY700; A, KYT06
(urrA),
N-OH-2-MeO-AAB did not induce such activity in any strains, even at 250/zM (data not shown).
Analysis of DNA damage in E. coil induced by N-OH-3-MeO-AAB and N-OH-2-MeO-AAB
The number of single-strand break or alkali-la- bile sites of DNA induced in N-OH-3-MeO-AAB- or N-OH-2-MeO-AAB-treated E. coil was esti- mated by sedimentation through alkaline sucrose gradient. E. coil AB2500 cells were exposed to 100/zM or 250/~M of the chemicals for 30 rain, and the spheroplasts from the treated cells were exposed to alkaline condition and applied to sedi- mentation analysis (Fig. 5). The average molecu- lar weight (Mw) of DNA from untreated cells corresponded to about 6.3 × 108 Da. After treat- ment with either N-OH-3-MeO-AAB or N-OH-2-
Fig. 5. Alkaline sedimentation profiles of DNA from E, coil AB2500 after treatment with N.OH-3-MeO-AAB or N-OH-2- MeO-AAB. Cells labeled with [methyl-'~H]thymine were tRated with 0 (o), 100 (e) or 250 (A)/ . tM of these azo dyes for 30 rain at 37°C. DNA fractions from alkaline sucrose gradient sedimentation were collected and precipitated by cold 5~ TCA with 100/~g of carrier BSA. The contents were filtered through a glass filter GF/C and subsequently washed, After being dried, radioactivities were assayed. The vertical bar indicates the position of [methyl-14C]thymidine-labelcd
~DNA (15 × 10 ~ Da).
69
MeO-AAB, DNA sedimented slowly, exhibiting a rough Mw of 3.5 × l0 s daitons at 100 /zM and 1.5 x l0 s daltons at 250 p.M of each chemical. Thus, the numbers of alkali-labile sites induced in the DNA by exposure to each chemical were similar, with approximately 1 or 3 breaks per cell at 100/zM or 250 p.M, respectively.
Using 32P-postlabeling analysis, we compared the DNA adducts formed in E. coli AB2500 treated with N-OH-3-MeO-AAB and N-OH-2- MeO-AAB. Five DNA adducts consisting of a major adduct (No. 1, 80% of total adducts) and 4 minor adducts (Nos. 2-5) were detected after the treatment with 100 /zM N-OH-3-MeO-AAB as shown in Fig. 6A. The total amount of DNA adducts was 8.5 adducts per 107 nucleotides. On the other hand, oniy one adduct (No. 1) could be detected after treatment with 100/zM of N-OH- 2-MeO-AAB, with an additional adduct (No. 2) being found to account for 8.5% of total adducts after treatment with 250/zM (Fig. 6B). The total amount of DNA adducts formed by N-OH-2- MeO-AAB was about 25-fold less than that by N-OH-3-MeO-AAB, namely, 3.5 or 8.1 adducts per 10 s nucleotides in DNA treated with 100 or 250/zM of N-OH-2-MeO-AAB, respectively.
Discussion
The present study clearly demonstrated that the cytotoxic and mutagenic effects of N-OH-3- MeO-AAB on E. coil strains were higher in uL'rA strains than in wild-type strains, suggesting that the UVR system in E. coil works to repair the DNA damage caused by N.OH-3-MeO-AAB. This assumption is also supported by the results indicating that induction of recA and umuC gene expression by N.OH-3-MeO-AAB was more ef- fective in uvrA strains (KY946 and KY706) than in wild-type strains (OE94 and KYT00). On the other hand, there was no appreciable difference in cytotoxic effects of N-OH-3.MeO-AAB be- tween the recA and the wild-type strains. The susceptibility of wild-type strains to 3-MeO-AAB is greater than to other carcinogens such as 4NQO. Then, the recA was found to be appar- ently not active on repair of 3-MeO-AAB dam- age. We suggested that the growth inhibition of E. coil wild-type cells by 3-MeO-AAB was due to
70
some nonspecific effects rather than to induced DNA damage.
Tarpley et al. (1982) reported that N-benzo- yloxy-N-methyl-4-aminoazobenzene (MAB) re- acted with DNA to cause the formation of apurinic/apyrimidinic sites. Therefore, in E. coil exposed to chemicals we analyzed, using alkaline sucrose gradient sedimentation, whether alkali- labile sites or single-stranded breaks were pro- duced in DNA. The sedimentation profiles indi- cated that both N-OH-3-MeO-AAB and N-OH- 2-MeO-AAB produced alkali-labile sites equally, 1-3 breaks per genome.
In this condition, however, N-OH-3-MeO-AAB reduced the percentage of cell survival to 43% and 8% at 100 #M and 250 #.M, respectively (data not shown), whereas over 80% of the cells survived after treatment with 250/~M of N-OH- 2-MeO-AAB. The conclusion that these alkali-la- bile sites might not contribute to the cytotoxic effects of these chemicals therefore appears war- ranted. The 32 P-postlabeling analysis revealed to- tal adduct levels induced by N-OH-2-MeO-AAB to be only 1/25 of these by N-OH-3-MeO.AAB.
Thus, the relative DNA binding potencies of N- hydroxyl derivatives of 3-MeO-AAB and 2-MeO- AAB in vivo correlate well with their cytotoxic potencies in E. coll.
Nevertheless, the differences in mutation fre- quency (over 300-fold difference) and SOS-induc- ing activity between N-OH-3-MeO-AAB and N- OH-2-MeO-AAB could not be explained solely on the basis of relative DNA binding potencies. N-OH-3-MeO-AAB formed one major adduct and 4 minor adducts, whereas N-OH-2-MeO- AAB formed only one major adduct correspond- ing to the 3-MeO-AAB. The minor adducts formed by N-OH-3-MeO-AAB corresponding to spots 2-5 in 32p-postlabeling TLC may be of importance in terms of mutagenicity and SOS induction.
Although the adducts formed by 3-MeO-AAB and 2-MeO-AAB have not yet been character- ized, C8-substituted deoxyguanosine may well be a major product (Hashimoto et al., 1981b; Kojima et al., 1991). As aminoazo-dye adducts, N-(de - oxyguanos in -8 -y l ) -methy l -4 -aminoazobenzene was identified as the major product formed by reac-
t, B
s
Fig, 6, Profiles of 3~P-labeled DNA adduets in E, colt ABZ500 exposed to 100 ~,M of N-OH-3-MeO-AAB (A) and 250 ~M of N-OH-2-MeO.AAB (B). 105 ~Ci or 210 ~.Ci of DNA digests from N-OH-3-MeO-AAB- or N-OH-2-MeO-AAB-treated cells, respectively, were applied to PEI sheets, ~2P-adduets were resolved by development in 3,5 M lithium formate, 8.5 M urea, pH 3.5 (bottom to top), followed by 0,8 M lithium chloride, 0,5 M Tris-HCl, 8.5 M urea, pH 8.0 (left to right). Autoradiography was
performed at - 8O°C for 6 h and 48 h for 3-MeO-AAB and 2-MeO-AAB DNA adducts, respectively.
tion of N-methyl-4-aminoazobenzene with DNA in vivo and in vitro (Lin et al., 1975; Beland et al., 1980).
In conclusion, our data indicate that biological differences between N-OH-3-MeO-AAB and N- OH-2-MeO-AAB may be reflections of the quali- tative and quantitative nature of their binding to DNA.
Acknowledgements
The authors are greatly indebted to Dr. K. Yamamoto for gifting E. coli strains. This work was supported in part by Grants-in-Aid for Can- cer Research from the Ministry of Education, Science and Culture of Japan.
References
Abelson, J,, and C.A. Thomas (1966) The anatomy of the T5 bacteriophage DNA molecule, J. Mol, Biol., 18, 262-291.
Bachmann, BJ. (1987) Derivations and genotypes of some mutant derivatives of Escherichia coil 8;-12, in: F.C. Neid- hardt, ],L, Ingraham, K.B. Low, B. Magasanik, M. Schaechter and H.E. Umbarger (Eds.), Escherichia coli and Salmonella typhimurium, Cellular and Molecular Biol- ogy, American Society for Microbiology, Washington, DC, pp. 1190-1219.
Beland, F.A., D.L. Tullis, F.F. Kadlubar, K.M. Straub and F.E, Evans (1980) Characterization of DNA adducts of the carcim)gen N-methyl-4-aminoazobenzene in vitro and in vivo, Chem.-Biol. Interact., 31, 1-17,
Ganesan, A,K., C,A, Smith ~nd A,A. Van Zeeland (1981) Measurement of the pynmidine dimer ccmtent of DNA in permeabilized bacterial or mammalian cells with endonu- clease V of bacteriophage T4, in: E,C, Friedberg and P,C, Hanawalt (Eds.), DNA Repair, Dekker, New York, pp. 89-9%
Gupta, R.C. (1985) Enhanced sensitivity of '~"P-postlabeling analysis of aromatic carcinogen: DNA adducts, Cancer Res., 45, 5656-5662,
Hashimoto, Y., H.K. Watanabe and M. Degawa (1981a) Mu- tagenicity of 4-aminoazobenzene, N.hydroxy-4-aminoazo- benzene, 4-nitrosoazobenzene, 4-nitroazobenzene, and their ring methoxylated derivatives in Salmonella, Gann, 72, 921-929.
Hashimoto, Y., M. Degawa, H,K. Watanabe and M. Tada (1981b) Amino acid conjugation of N.hydroxy-4-
71
aminoazobenzene dyes: a possible activation process of carcinogenic 4-aminoazobenzene dyes to the ultimate mu- tagenic or carcinogenic metabolites, Gann, 72, 937-943.
Kojima, M., M. Degawa, Y. Hashimoto and M. Tada (1991) 32p-Postlabeling analysis of DNA adducts in rat livers after treatment with genotoxic and non-genotoxic 4- aminoazobenzene derivatives, Cancer Lett., 58, 199-204.
Kondo, S., H. ichikawa, K. Iwo and T. Kato (1970) Base- change mutagenesis and prophagc induction in strains of Escherichia coli with different DNA repair capacities, Ge- netics, 66, 187-217.
Lin, J.K., B. Schmall, I.D. Sharpe, I. Miura, J.A. Miller and E.C. Miller (1975) N-substitution of carbon 8 in guanosine and deoxyguanosine by the carcinogen N-benzoyloxy-N- methyl-4-aminoazobenzene in vitro, Cancer Res., 35, 832- 843.
Marmur, J. (19611 A procedure for the isolation of deoxyri- bonucleic acid from micro-org:misms, J. Mol. Biol., 3, 208-218.
Miller, J.A., and E.C. Miller (19611 The carcinogenicity of 3-methoxy-4-aminoazobenzene and its N-methyl deriva- tives for extrahepatic tissues of the rat. Cancer Res., 21, 1068-1072.
Miller, J.H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Odashima, S., and Y. Hashimoto (1968) Carcinogenicity and target organs of methoxyl derivatives of 4-aminoazobcn- zene in rats. I. 3-Methoxy- and 3,4'-dimethoxy-4- aminoazobenzene, Gann, 59, 131-143.
Odashima, S., and Y. Hashimoto (19701 Carcinogenieity and target organs of methoxyl derivatives of 4-aminoazoben- zene in rats. II. Effect of various concentrations of 3- methoxy- and 2,5-dimethoxy-4-aminoazobenzene in the diet, Gann, 61,153-16(l.
Tarpley, W.G., J.A. Miller and E.C. Miller (1982) Rapid release of carcinogen-guanine adducts from DNA after reaction with N-acetoxy-2-acetylaminofluorene or N-ben- zoyloxy-N-methyl-4-aminoazohenzene, Carcinogenesis, 3, 81-88.
Vogel, HJ., and D,M. Bonner (1956) Acetylornithinase of E. coil: partial purification and some properties, J, Biol, Chem., 218, 97-106,
Watanabe, H.K., and Y. Hashimoto (19811 Unscheduled DNA synthesis induced by 4-aminoazobenzene, N-hydroxy-4- aminoazobenzene, and their derivatives in primary cul- tures of rat and mouse hepatocytes, Gann, 72, 930-936.
Yamamoto, K., T. Higashikawa, K. Ohta and Y. Oda (1985) A loss of urrA function decreases the induction of the SOS functions recA and umuC by mitomycin C in Escherichia coli, Mutation Res., 149, 297-302.
