N2 Atom of Guanine and N6 Atom of Adenine Residues As Sites for Covalent Binding of Metabolically Activated 1'-Hydroxysafrole to Mouse Liver DMA in V/Vo1

Home , Deoxyguanosine

[CANCER RESEARCH 41, 2664-2671, July 1981] 0008-5472/81 /0041-OOOOS02.00 N2 Atom of Guanine and N6 Atom of Adenine Residues as Sites for Covalent Binding of Metabolically Activated 1'-Hydroxysafrole to Mouse Liver DMA in V/vo1

David H. Phillips,2 James A. Miller,3 Elizabeth C. Miller, and Bruce Adams

McArdle Laboratory for Cancer Research [D. H. P., J. A. M., E. C. M.Â¡and Department of Chemistry Â¡B.A.], University of Wisconsin, Madison. Wisconsin 53706

ABSTRACT use of safrole and sassafras oil (containing safrole) as food Administration of 1'-[2',3'-3H]hydroxysafrole to adult female additives was banned in the United States in 1960, safrole continues to be ingested in small amounts by humans since it mice resulted in the formation of DMA-, ribosomal RNA-, and is a minor component of a number of other essential oils and of protein-bound adducts in the liver that reached maximum levels some herbs and spices, including anise, basil, nutmeg, mace, within 24 hr. The levels of all three macromolecule-bound and pepper (12, 13, 20). adducts decreased rapidly between 1 and 3 days after injec Current evidence (reviewed in Refs. 22 and 27) suggests tion, at which time the amounts of the DMA-bound adducts that a property common to most chemical carcinogens is that essentially plateaued at approximately 15% of the maximum their biologically active forms are electrophilic. The majority of level. The amounts of the protein and ribosomal RNA adducts carcinogens are not electrophilic as such and must undergo were very low by 20 days. metabolic activation in vivo to reactive species of this type. Comparison by high-performance liquid chromatography of Studies on safrole (6, 7, 24, 28-30, 34, 35) indicated the the deoxyribonucleoside adducts obtained from the hepatic metabolic formation of several electrophilic metabolites. Borch- DMA with those formed by reaction of deoxyguanosine and ert ef al. (5, 6) showed that 1'-hydroxysafrole is a major deoxyadenosine with 1'-acetoxysafrole, 1'-hydroxysafrole- 2',3'-oxide, and 1'-oxosafrole indicated that the four in vivo metabolite of safrole in the rat and mouse and that it possesses greater carcinogenic activity than does the parent compound adducts studied were derived from an ester of 1'-hydroxysaf- in these species. 1'-Hydroxysafrole undergoes further metab role. Three of the four in vivo adducts comigrated with adducts olism by rat and mouse liver preparations to 1'-sulfonoxysafrole formed by reaction of 1'-acetoxysafrole with deoxyguanosine; and 1'-hydroxysafrole-2',3'-oxide (30, 34) (Chart 1). Both the the fourth adduct comigrated with the major product of the latter metabolite and a model analog of the former, 1'-acetox reaction of this ester with deoxyadenosine. Adduct formation ysafrole, possess electrophilic, carcinogenic, and mutagenic in vivo at low levels by the other two electrophilic metabolites activity (5, 6, 29, 34, 35). In addition, 1'-oxosafrole, small was not excluded. The three adducts obtained by reaction of amounts of which are found as Mannich base derivatives in the 1'-acetoxysafrole with deoxyguanosine appeared to be substi urine of rats administered safrole (24), exhibits electrophilic tuted on the 2-amino group of the guanine residue on the basis activity (34). However, 1'-oxosafrole has not shown mutagenic of their partitions between aqueous buffer solutions and 1- or carcinogenic activity (34, 35). We have reported recently butanohethyl ether as a function of pH and their retention of 3H (26) that 1'-hydroxyestragole, a compound structurally related from [8-3H]deoxyguanosine. The corresponding three adducts to 1'-hydroxysafrole and a proximate carcinogenic metabolite derived from the hepatic DNA of mice given 1'-[2',3'-3H]hy- of the natural flavoring agent estragÃ³le (1-allyl-4-methoxyben- droxysafrole had pH partition patterns not significantly different zene) (9), is metabolically activated in mouse liver in vivo to a from the three adducts formed in vitro. Adduct II was further derivative, presumably a 1'-ester, that reacts covalently with characterized from its nuclear magnetic resonance spectrum the exocyclic amino groups of guanine and adenine residues as /V2-(frans-isosafrol-3'-yl)deoxyguanosine. Adduct IV, de in DNA. The present paper reports that the DNA adducts rived from the reaction of 1'-acetoxysafrole with deoxyadeno formed in mouse liver after administration of 1'-hydroxysafrole sine 5'-phosphate, was characterized in the same manner as are analogous to the 1'-hydroxyestragole:DNA adducts (26), iVXfrans-isosafrol-S'-yOdeoxyadenosine. and evidence for their structures is presented. The earlier assignment of the structure of the major product of the reaction INTRODUCTION of 1'-acetoxysafrole with GMP as O6-(isosafrol-3'-yl)guanylic acid (6, 34) was found to be incorrect. Safrole [1 -allyl-3,4(methylenedioxy)benzene], a naturally oc curring flavoring agent that is the major constituent of oil of MATERIALS AND METHODS sassafras (12, 13, 20), possesses weak hepatocarcinogenic activity when fed to adult rats or mice (1, 5, 14, 15, 21) and Safrole Derivatives. The syntheses of 1'-hydroxysafrole and moderate hepatocarcinogenic activity when injected into mice 1'-acetoxysafrole (6) and of 1'-oxosafrole, 1'-hydroxysafrole- during the first few weeks after birth (5, 11, 35). Although the 2',3'-oxide, 1'-hydroxy-2',3'-dehydrosafrole, and 1'-{2',3'- 3H]hydroxysafrole (34) have been described previously. ' This work was supported by Grants CA-07175 and CA-22484 from the National Cancer Institute. USPHS. Hepatic DNA, rRNA, and Protein from Mice Treated with 2 Present address: Department of Biological Sciences, Stanford University. 1'-{3H]Hydroxysafrole. Female CD-1 mice (Charles River Stanford, Calif. 94305. 3 To whom requests for reprints should be addressed. Breeding Laboratory, Wilmington, Mass.), 8 to 10 weeks old (mean weight, 30 g) were given i.p. injections of 1'-[2',3'- Received December 5. 1980; accepted April 1, 1981.

2664 CANCER RESEARCH VOL. 41

1-Ã‡H- incubated with 4 units of alkaline phosphatase (Sigma) at 37Â° for 48 hr. The nucleosides which precipitated were washed with water and dried in a vacuum. HPLC analysis revealed that H-C-CH =CH2 the material (18 mg) obtained from Fractions 51 to 55 from the OH I-HYDROXYSAFROLE dGMP reaction consisted of a single product. Earlier fractions contained a mixture of this product and unreacted dGuo, and later fractions contained 2 minor products in addition to the major product found in Fractions 51 to 55. Similarly, the material (28 mg) from Fractions 61 to 70 of the dAMP reaction -,-CH iâ€”CHo -CH, mixture contained a single product; earlier fractions contained additional unreacted dAdo, and subsequent fractions con tained additional adducts. C-CH=CH2 H-Ã‡-CH=CH2 H-Ã‡-CH-CH2 NMR spectra were determined on solutions of the adducts in 0 0-S03H OH M dimethyl sulfoxide-d6 (5 mg in 0.4 ml) by use of a 270-MHz I'-OXO- I'-SULFONOXY- I-HYDROXY Bruker WH 270 spectrometer equipped with a B-NC 12 Nicolet SAFROLE SAFROLE SAFROLE- 2',y- OXIDE computer. Chart 1. Pathways of metabolism of 1'-hydroxysafrole, a proximate carcino HPLC Chromatography. Chromatography of DNA hydroly genic metabolite of safrole, to electrophilic species. sates and marker nucleoside adducts was performed on an ALC/GPC 294 liquid Chromatograph (Waters Associates, Mil- 3H]hydroxysafrole (404 mCi/mmol, 12 jumol/mouse in 0.1 ml ford, Mass.) equipped with a Model U6K injector, a Model 660 trioctanoin). At the times indicated, the animals were killed by solvent programmer, a Model 440 absorbance detector, an cervical dislocation, and the livers from groups of 5 mice were Omniscribe B-5000 strip chart recorder (Houston Instruments, pooled for isolation of DMA, rRNA, and protein by the method Austin, Texas), and a Spherisorb ODS 5 /im reverse-phase of Irving and Veazey (16). DNA and rRNA were dissolved in column (Altex Scientific Inc., Berkeley, Calif.). The solvent Tris buffer (0.01 M, pH 7.0) and hydrolyzed enzymatically to systems used were: System A, 100% water for 5 min followed nucleosides by the method of Baird and Brookes (2). Protein by a linear (Program 6) gradient of 15 to 40% acetonitrile:water was dissolved in Soluene (Packard Instrument Co., Inc., Down for 35 min; System B, 25% acetonitrile:water; System C, 45% ers Grove, III.). Aliquots of digests of the 3 macromolecules methanol:water. For each system, the flow rate used was 2 ml/ were added to Aquassure scintillation fluid (New England Nu min. As noted previously (26), the retention times of adducts clear, Boston, Mass.) and assayed for radioactivity in Isocap/ may be altered by 1 to 2 min depending on the quantity of 300 (Nuclear Chicago, Inc., Chicago, III.) or Mark Ml/6880 other materials, such as unmodified nucleosides, necessarily (Searle Analytic, Inc., Des Plaines, III.) scintillation spectrome coinjected with them. Thus, the retention times quoted for ters. External standardization was used to convert all of the adducts are not absolute, and it was essential for determining data to dpm. the identical natures of in vivo- and in wfro-derived adducts Preparation of Nucleoside Adducts. 14C-Labeled nucleo- that they be eluted simultaneously after coinjection on HPLC. side adduct markers were prepared by reacting the appropriate pH Partition Coefficient Patterns of Adducts. Aliquots (50 1'-hydroxysafrole derivative (10 mg in 0.5 ml of ethanol) with to 100 ÃÃl)ofthe adduct-containing fractions eluted from HPLC [14C]dGuo4 or [14C]dAdo (2 mg containing 0.5 jiCi in 0.5 ml columns were partitioned by the procedure of Moore and 0.01 M Tris, pH 7.0). Unlabeled nucleosides were obtained Koreeda (23) between 1 ml of 0.05 M buffer solutions (pH from Sigma Chemical Co., St. Louis, Mo.; [8-'"C]dGuo and [8- range, 1 to 13) and 1 ml of 1-butanol:ethyl ether (20:80), each 14C]dAdo were purchased from Schwarz/Mann, Orangeburg, of which had been presaturated with the other. After vigorous N. Y., and New England Nuclear, respectively. Aliquots (50 jxl) shaking, the phases were separated, and portions (0.7 ml) of of the reaction mixtures were coinjected on HPLC with 3H- each phase were assayed for radioactivity. labeled DNA hydrolysates from mouse liver. In some experi ments, 1'-acetoxysafrole (10 mg in 0.5 ml ethanol) was reacted RESULTS with dGuo (2 mg in 0.5 ml 0.01 M Tris, pH 7.0) containing 5 juCi [8-3H]dGuo (Schwarz/Mann) and 0.5 jtiCi [8-14C]dGuo. Binding of 1-Hydroxysafrole to Hepatic Macromolecules. For large-scale preparation of 1'-acetoxysafroleinucleoside The levels to which 1'-hydroxysafrole was bound to liver mac adducts, 1'-acetoxysafrole (2.3 g) in ethanol (100 ml) was romolecules after administration of a single dose i.p. to adult reacted with either dGMP or dAMP (1 g) in 75 ml 0.01 M Tris female CD-1 mice are shown in Chart 2. Maximum binding to buffer (pH 7.0) at 37Â° for 17 hr with gentle shaking. The protein, rRNA, and DNA occurred at 5 hr, and by 3 days after reaction mixture was then extracted with ethyl ether (4 x 200 treatment, the amount of radioactivity bound to all 3 macro- ml). The aqueous phase was reduced in volume to 5 ml and molecules had fallen by 85%. The levels of binding to protein applied to a Sephadex LH-20 column (50 x 3 cm) that was and rRNA continued to fall up to 20 days after treatment, but eluted with water, and fractions (approximately 8 ml) were the level of binding to DNA remained relatively constant be collected. Pools of 5 to 10 fractions were freeze dried, and the tween 3 and 20 days after treatment. residues were dissolved in 3 ml 0.1 M Tris buffer, (pH 9.0) and HPLC Analysis of the Hepatic DNA Hydrolysates. The elution profiles from reverse-phase HPLC for hepatic DNA hydrol ysates from mice killed 23 hr after a single i.p. injection of 1'- 4 The abbreviations used are: dGuo, deoxyguanosine; dAdo. deoxyadenosine; [2',3'-3H]hydroxysafrole are shown in Charts 3 to 5. A signifi HPLC, high-performance liquid chromatography; NMR, nuclear magnetic reso nance. cant amount of radioactivity eluted in the region of the chro-

JULY1981 2665

300, Adducts I to IV. However, some of these products eluted in the intermediate region (20 lo 24 min) of Ihe chromalograph lhal conlained low levels of 3H-labeled material (Chart 5A). The oxosafrole produci al 23.5 min consislenlly moved about 0.5 min faster than did Adduct I in Solvent A. Reaction of 1'- oxosafrole wilh [14C]dAdo gave a major adduci which eluled close lo, bul noi simullaneously wilh, the in vivo Adduci II in bolh Solvenl Systems A (Chart 5B) and C (dala noi shown). The coelution from HPLC columns of the in vivo producÃs wilh those formed in vitro from 1'-aceloxysafrole indicated lhal 1'-hydroxysafrole binds lo mouse liver DNA principally via an eleclrophilic ester. This cochromalography also indicated thai adducts were formed wilh bolh Ihe guanine (Adducts I, II, and III) and adenine (Adduct IV) bases in DNA. However, the pos sible formation in vivo of minor adducls from reaction of DNA with 1'-hydroxysafrole-2',3'-oxide or 1'-oxosafrole was noi excluded. The 3H in mouse liver DNA hydrolysales lhal eluted from HPLC columns at 15 lo 22 min may be accounted for by such adducts. Alternately, Ihese low levels of 3H may denote

l 2 3 4 5 IO polar degradation producÃs of Adducls I lo IV, which resulted DAYS AFTER TREATMENT from chemical or enzymalic modification of the methylenedioxy Chart 2. Binding of 1'-hydroxysafrole to the liver protein, rRNA, and DMA of ring of Ihe safrole moiety. groups of 5 adult female CD-1 mice following treatment i.p. with 12 /imol/30-g pH Partition Coefficient Patterns of Adducts Obtained in mouse. Two experiments per time point were carried out. Bars, S.D. Vitro and in Vivo. To further investigate Ihe nalure of Ihe dGuo Adducls I to III, both the adducls obtained in vivo and those matographs where unmodified nucleosides eluted (9 to 12 from the reaction of 1'-acetoxysafrole wilh [14C]dGuo were min). In the intermediate regions of the chromatographs, from collected from HPLC columns and partilioned between 15 to 22 min, low levels of radioactivity eluted, but no well- aqueous buffer solutions and 1-bulanol:elhyl elher (20:80) defined peaks were discernible. In the later regions of the chromatographs, 4 well-defined peaks of radioactivity eluted; 1-Â®MOUSE-LIVER > 1 1 these were designated I to IV in order of elution. HPLC analysis 4000i of the DNA hydrolysates showed essentially the same profiles IILIf:DNA HYDROLYSATE at all time points. The radioactivity in the later regions of the 200002000Eiu chromatographs was distributed between the 4 adducts in the iJLri^^ following proportions: I, 33 Â±6% (S.D.); II, 62 Â±5%, III, 3 Â± udGuo 11 L m 1%, and IV, 2 Â± 1%. Although some variation in the relative amounts of Adducts I to IV was observed, the differences were .nL Â«pCJOEOXYGUANOSINE F-I*-ACETOXYSAFROLE random with respect to the time at which the mice were killed. Coinjection on HPLC of hepatic DNA hydrolysates and an aliquot of the reaction mixture of 1'-acetoxysafrole with 100004000o.â€¢o,1 [14C]dGuo (resulted in 14C-labeled products that coeluted in fL^:MOUSE-LIVER Solvent System A with the 3H-labeled in vivo Adducts I, II, and III (Chart 3A). The in vivo Adduci IV coeluted with a '4C-labeled product from the reaction of 1'-acetoxysafrole with [14C]dAdo (Chart 3B). Comigration of Adducts I, II, and III with 1'-acelox- iil" DNA HYDROLYSATE ysafrole:dGuo adducts and of Adduct IV with a 1'-aceloxysaf- 20000.'800Eo.T>Â£ rJv\ n role:dAdo adduci was also observed on HPLC with Solvent Systems B and C. Representalive relenlion limes were as EdAdo ^J ^ l^__ follows: Solvent Syslem B: I, 9.0 min; II, 11.0 min; III, 13.5 min; and IV, 21.5 min; Syslem C: I, 9.5 min; II, 12.5 min; III, 15 min; .- r f-ACETOXYSAFROLE â€¢¿[^CJDEOXYADENOSINE and IV, 23.5 min. When 1'-hydroxysafrole-2',3'-oxide was reacled wilh [14C]- 400Oâ€¢nvL^Jl*-W dGuo, a single major adduci was formed thai eluted al 13.5 min on HPLC wilh Solvenl System A. Low levels of 3H-labeled malerial from liver DNA hydrolysales also eluled in Ihis general . n . iL region (Chart 4/\). Some of Ihe producÃs of Ihe reaclion of 1'- IO 15 20 25 30 35 40 hydroxysafrole-2',3'-oxide with [14C]dAdo also eluted with in MINUTES Chart 3. HPLC profiles of DNA hydrolysates and marker nucleosides. DNA termediate relenlion times lhal were less lhan those of the in hydrolysates from the livers of mice given injections of 1'-{3H]hydroxysafrole vivo Adducts I to IV (Chart 48). The reaclion of 1'-oxosafrole were cochromatographed with aliquots of the reaction mixture of 1'-acetoxysaf with [14C]dGuo gave a complex pattern of producÃs, none of role reacted with ["CJdGuo W and ("CJdAdo (B). The eluting solvent was: 0 to which coeluted exaclly wilh any of Ine 3H-labeled in vivo 5 min, 100% water; 5 to 40 min, 15 to 40% acetonitrile:water (linear gradient) at a flow rate of 2 ml/min.

2666 CANCER RESEARCH VOL. 41

dGuo derivatives) of each peak by HPLC with the solvent 40001I HYDROLYSATEInIVDNA systems used. Structure of the dGuo Adduct II. The major nucleoside adduct derived from the large-scale reaction of 1'-acetoxysaf 20008000io role with dGMP (see "Materials and Methods") comigrated on H_^â€” HPLC in Solvent Systems B and C (see "Materials and Meth n.dGuo--l'-HYDROXYSAFROLE-2',3'-OXIDE1-Â»[I4CJDEOXYGUANOSINE-Â®MOUSE-LIVER-~-* L-~n nr1 LT!n--1m 3Â» ods") with the major in vivo Adduct II. Its proton NMR spectrum, shown in Chart 7, was not that of a safrole derivative. Instead, the NMR spectrum is indicative of a 3'-substituted Â¡sosafrole

40004000Eo.T>in [1 -propenyl-3,4(methylenedioxy)benzene] derivative with frans configuration (Jr,2. = 16.0 Hz) at the C-1', C-2' double bond. The signal for the A/2 proton appeared as a triplet at 6.67 ppm and integrated as a single proton. Its coupling to the C-3' protons (4.04 ppm) was demonstrated by its changing to a HYDROLYSATEIdAdoP-â€¢-j^U-r^iuDNA singlet upon irradiation of the C-3' signal. The N7 signal dis appeared entirely when the sample was exchanged with D20, 200Â°0BOOB,â€¢oÂ« and at the same time, the C-3' signal changed to a doublet. _.vI'-HYDROXYSAFROLE-^.J-OXIDEÂ»[l4CJDEOXYADENOSINE-VsThe C-8 proton of guanine appeared at 7.90 ppm, which was similar to its position in the spectrum of dGuo; it did not exchange with D2O. On the basis of these findings, Adduct II is unequivocally assigned the structure A/2-(frans-isosafrol-3'- yl)deoxyguanosine. The N-1 proton signal was not observed. 400OÂ®MOUSE-LIVER This result is at variance with the assignment reported pre liWL. r^ - viously of the major product of the reaction of 1'-acetoxysafrole with GMP as O6-(isosafrol-3'-yl)guanylic acid on the basis of 15 20 25 30 35 40 its instability under mildly acidic conditions (34) and a 60-MHz MINUTES proton NMR spectrum of a major acetylated nucleoside adduct Chart 4. HPLC profiles of DNA hydrolysates and marker nucleosides. DNA hydrolysates from the livers of mice given injections of 1'-[3H]hydroxysafrole derived from this material (6). Therefore, we have reexamined were cochromatographed with aliquots of the reaction mixture of 1'-hydroxysaf- some of the original material (6). Since none of the acetylated role-2',3'-oxide reacted with [14C]dGuo (A) and ["CJdAdo (ÃŸ).The eluting solvent was as described in the legend to Chart 2.

4000 (Chart 6). For Adducts I and II, the close agreement between MOUSE-LIVER DNA HYDROLYSATE the partition coefficients of the in vivo- and in w'fro-derived samples of each adduci provided further evidence for their I 2000 being identical. The greater variation in the patterns of Adduct III formed in vivo or in vitro was probably a consequence of the JÃŒ dGuo I'-OXOSAFROLE â€¢¿[I4c]DEOXYGUANOSINE much smaller amounts of radioactive material (approximately 2000 100 dpm/experimental point) with which the experiment was conducted. However, the data clearly indicated pK0 values at both acidic and basic pH for each of Adducts I to III. These 1000 findings excluded the possibility that any of the adducts were O6- or N-1-substituted dGuo derivatives, since those adducts n. /lA would lack an ionizable proton at the N-1 position of guanine 4000 and hence not show a pKa at basic pH. The patterns shown in MOUSE-LIVER DNA HYDROLYSATE Chart 6 indicate that substitution occurred at the C-8, N-7, or Q. N! position of guanine. 2000 Reaction of 1'-Acetoxysafrole with [8-14C, 8-3H]dGuo. When 1'-acetoxysafrole was reacted with dGuo labeled at E position 8 of guanine with 14C and 3H and the products were dAdo f-OXOSAFROLE â€¢¿[I4CJDEOXYADENOSINE separated by HPLC, Adduct II retained 97% of the tritium 8000 (Table 1). Since loss of tritium from position 8 of guanine would be expected for both C-8 and N-7-substituted dGuo derivatives 4000 (32), this and the above results indicate that Adduct II contains a guanine residue substituted at the 2-amino group. Adducts I ^a and III also showed a high percentage of tritium retention, but 20 25 30 35 40 the levels were consistently lower than those for dGuo or MINUTES Adduct II (Table 1, 86 and 82%, respectively). It is therefore Chart 5. HPLC profiles of DNA hydrolysates and marker nucleosides. DNA possible that these in w'fro-derived adduci peaks contain minor hydrolysates from the livers of mice given injections of 1'-pH]hydroxysafrole were cochromatographed with aliquots of the reaction mixture of 1'-oxosafrole amounts (<20%) of C-8 and/or N-7 adducts, which were not reacted with ["CJdGuo (A) and ['4C)dAdo (8). The eluting solvent was as separable from the major components (i.e., A/2-substituted described in the legend to Chart 2.

JULY 1981 2667

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 1981 American Association for Cancer Research. D. H. Phillips et al. adduci was available, the 1'-acetoxysafrole:GMP adduct from the absorbance at 254 nm was eluted in a single peak; 2 less the latter study was converted to the nucleoside adduct, as polar minor peaks accounted for the remaining 15%. The pH described for the dGMP and dAMP adducts in "Materials and partition coefficient pattern of the adduct, determined as de Methods." HPLC analysis of the material indicated that 85% of scribed in "Materials and Methods" except that absorbance at 254 nm was used to determine the amount of adduct in each phase, showed pKa values at both acidic (2.2) and basic (9.8) pH. This finding excludes the possibility of the adduct being an 80 O6-substituted derivative, since such an adduct would lack an ionizable proton at the N-1 position and thus not show a basic 60 pKa. The 270-MHz proton NMR spectrum of the 1'-acetoxysaf- 40 role:Guo adduct was similar to that of Adduct II (Chart 7) except 20- that the former was characteristic of a ribonucleoside, rather than of a deoxyribonucleoside. Thus, the salient features were that the N2 proton signal was a triplet at 6.71 ppm and inte 0 ADDUCT I grated as a single proton and that the N-1 proton signal was 80 seen as an extremely broad signal at 9.37 ppm. The parts of the spectrum due to the safrole substituent were indicative of 60 a 3'-substituted frans-isosafrole derivative. On the basis of this reexamination of the 1'-acetoxysafrole:Guo adduct with im 40 proved techniques, it appears that it is N2-(frans-isosafrol-3'- yOguanosine. 20 We have also reexamined the original mass and NMR spectra of the acetylated adduct that was analyzed previously (6). The 0 mass spectrum clearly shows a peak at m/e = 611, which is

60 consistent with a tetraacetate derivative and which is the high est mass peak observed. However, the integral of the signal 60- Table 1 40- 3H:"C ratios of products of the reaction of 1'-acetoxysafrole with [8-"C; 8-3HldGuo 1'-Acetoxysafrole (10 mg in 0.5 ml of ethanol) was reacted with dGuo (2 mg 20- in 0.5 ml Tris, 0.01 M, pH 7.0) containing 5 /iCi of [8-3H]dGuo and 0.5 fiCi of [8- MC]dGuo. A 0.1-ml aliquot was chromatographed on HPLC, and the fractions were assayed for 3H and "C. 2 4 6 8 10 12 of retention of3H100 pH dGuo Chart 6. Partition of dGuo adducts between 0.05 M aqueous buffers, pH 1 to 13 (23) and 1-butanohethyl ether (20:80). â€¢¿,3H-labeledadducts derived in vivo; Adduct I 7.98.9 86 O, "C-labeled adducts derived in vitro from the reaction of 1'-acetoxysafrole with Adduct II 97 ("C]dGuo. Adduct III3H:'4C9.2 7.4% 82

DMSO

ADDUCT

Chart 7. 'H NMR spectrum (270 MHz) of Adduct II in dimethyl sulfoxide (DMSO)-de (5 mg in 0.4 ml). The ma terial was obtained by treatment with alkaline phosphatase of the major product of the reaction of 1'-acetoxysafrole with dGMP (see "Materials and Methods").

2668 CANCER RESEARCH VOL. 41

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 1981 American Association for Cancer Research. 1'-Hydroxysafrole-DNA Adducts in Mouse Liver for the acetyl protons in the 60-MHz NMR spectrum accounts 1',2'-dihydroisosafrole. Adduct III was available in such small for about 10 protons, i.e., intermediate between that expected amounts that further studies were not undertaken. for triacetate (9 protons) and tetraacetate (12 protons) deriva tives. A broad signal at 7.9 ppm (the C-8 signal appeared at DISCUSSION 7.65 ppm), which was exchangeable with D2O, was not as signed in the original study (6). It is possible that this signal is The pattern of DMA adducts formed in the livers of mice due to the N2 proton and that the signal at 12.35 ppm, originally administered 1'-hydroxysafrole is similar to that which we have assigned to the A/2 proton, is due to the N-1 proton. In view of described recently for the closely related compound 1'-hydrox the fact that several A/2-substituted dGuo or guanosine adducts yestragole (26). In each case, 3 dGuo and one dAdo adducts derived from other chemical carcinogens are hydrolyzed to were formed, and all involved covalent attachment of the car various extents under acidic conditions (17, 18, 26), it is now cinogen moiety to the exocyclic amino groups of the guanine evident that ease of hydrolysis by acid is not a reliable indicator and adenine residues. The major adduct (II) formed by each of an O6-substituted guanine derivative (34). compound contains the 2-amino group of guanine attached to Structure of the dAdo Adduci IV. The nucleoside adduct position 3' of frans-isoestragole or frans-isosafrole. Although derived from the large-scale reaction of 1'-acetoxysafrole with the 2 other dGuo adducts (I and III) formed by 1'-hydroxysafrole dAMP (see "Materials and Methods") comigrated on HPLC in have not been fully characterized, it is probable from their Solvent Systems B and C with the minor in vivo Adduct IV. Its retention times on HPLC relative to that of Adduct II that they NMR spectrum is shown in Chart 8. As in the case of Adduct are analogous to the 1'-hydroxyestragole Adducts I and III. II (Chart 7), the C-1 ', C-2', and C-3' signals are indicative of a Evidence was presented that the structures of the latter 2 frans-isosafrole derivative (Jr.2. = 16.0 Hz) substituted at the adducts were A/2-(estragol-1 '-yOdeoxyguanosine and A/2-(c/s- C-3' position. The signal at 8.06 ppm was exchangeable with isoestragol-3'-yl)deoxyguanosine, respectively (26). A minor D2O and is assigned to the 6-amino group. The signal inte adduct (IV) was also formed from both 1'-hydroxysafrole and grated as a single proton and exhibited pronounced broaden 1'-hydroxyestragole; these adducts involved conjugation of the ing, as did the Â¡sosafrole C-3' signal (4.23 ppm) and the 6-amino group of adenine residues to position 3' of frans- adenine C-2 signal (8.21 ppm). The reason for these broad- isosafrole or frans-isoestragole. The 3'-substituted isosafrole enings is not clear, but, as discussed earlier (26), similar adducts (II and IV) could arise by an SN1 mechanism, in which broadenings have been observed in the spectrum of the refer the purine bases of DNA substituted position 3' of the carbo- ence compound A/6-methyladenosine and of other A/6-substi- nium ion generated by the loss of the 1'-ester group. Alterna tuted dAdo adducts (1 7, 26). Thus, Adduct IV is assigned the tively, they could arise by an SN2' mechanism, whereby the structure A/6-(frans-isosafrol-3'-yl)deoxyadenosine. 2',3'-allylic double bond shifted to become a 1',2'-p,">penylic Structures of Adducts I and III. As noted above, substitution double bond with loss of the ester group. These pocsible of Adducts I and III apparently occurred on the N2 atom of the reaction mechanisms were discussed in more detail in >:"" guanine residues. An attempt was made to determine the earlier paper on the adducts formed from 1'-acetoxyestragole position of attachment to the safrole residue in Adduct I by (26). reduction with H2:platinum, acid hydrolysis of the reduced Although the major share of the 3H in the hydrolysates of adduct, and identification of the resulting dihydrosafrole deriv hepatic DNA from mice given either 1'-[3H]hydroxysafrole or ative; this procedure was successful in the previous study on 1'-[3H]hydroxyestragole migrates with Adducts I to IV formed a DMA adduct formed from 1'-hydroxyestragole (26). However, from either 1'-acetoxysafrole or 1'-acetoxyestragole and dGuo this approach was not practical in the present study because or dAdo, some 3H products eluted prior to these adducts of our inability to resolve adequately on HPLC the reference (Charts 3 to 5; Ref. 26). Furthermore, the amounts of these 3H compounds 1'-hydroxy-2',3'-dihydrosafrole and 3'-hydroxy- products, which were eluted between 7 and 20 min, appeared

ADDUCI 12

-OCH20- H20 ->-H C-8

Chart 8. 'H NMR spectrum (270 MHz) of Adduct IV in dimethyl sulfoxide (DMSOWe (5 mg in 0.4 ml). The material was obtained by treatment with alkaline phosphatase of the major product of the reaction of 1'-acetoxysafrole with dAMP (see "Materials and Methods").

C:Ã

3-OH S'-OH ta UvJ M

JULY 1981 2669

to be greater when 1'-[3H]hydroxysafrole was administered. a proximate carcinogenic metabolite of safrole in the rat and mouse. Cancer Res., 33. 590-600, 1973. The reasons for this quantitative difference are not known. 6. Borchert, P., Wislocki, P. G., Miller, J. A., and Miller, E. C. The metabolism Some products from the reaction of 1'-hydroxysafrole-2',3'- of the naturally occurring hepatocarcinogen safrole to 1'-hydroxysafrole and oxide and 1'-oxosafrole with dGuo or dAdo eluted in this area, the electrophilic reactivity of 1'-acetoxysafrole. Cancer Res., 33: 575-589, 1973. and it is possible that these electrophiles, but not the corre 7. Delaforge, M., Janiaud. P., Chessebeuf, M., Padieu. P.. and Maume. B. F. sponding electrophilic derivatives from 1'-hydroxyestragole, Possible existence of the epoxide-diol metabolic pathway for nepatocarcin- contributed small but significant amounts to the DMA adducts ogenic safrole in cultured rat liver cells, as compared with whole animal: a metabolic study by mass spectrometry. In: A. Frigerio and N. Castagnoli formed in vivo. Some radioactivity from tritiated polycyclic (eds.). Advances in Mass Spectrometry in Biochemistry and Medicine, Vol. aromatic hydrocarbons has been found to elute very early on 2, pp. 65-89. New York: Spectrum Publications, 1976. chromatography of cellular DNA hydrolysates (2, 3, 10, 25). 8. Dipple, A., Hayes, M. E., and Constantino, N. Response of DMBA-DNA adducts to excision processes in mouse embryo cells and to enzymatic Some of this radioactivity has been attributed to incorporation digestion in vitro. Proc. Am. Assoc. Cancer Res., 27: 117, 1980. 9. Drinkwater, N. R., Miller, E. C., Miller, J. A., and Pilot, H. C. Hepatocarcin- of tritium into normal deoxyribonucleosides (2, 3, 25). How ogenicity of estragÃ³le (1-allyl-4-methoxybenzene) and 1'-hydroxyestragole ever, in one case, its dependence on the enzymatic method in the mouse and mutagenicity of 1'-acetoxyestragole in bacteria. J. Nati. used for the hydrolysis of the DNA suggested that there might Cancer Inst., 57: 1323-1331, 1976. also be carcinogen-modified fragments of DNA resistant to 10. Eastman, A., and Bresnick, E. Persistent binding of 3-methylcholanthrene to mouse lung DNA and its correlation with susceptibility to pulmonary neopla complete hydrolysis (8). Whether or not such circumstances sia. Cancer Res., 39: 2400-2405, 1979. are involved in the substantial amounts of 3H that eluted be 11. Epstein, S. S., Fujii, K., Andrea, J., and Mantel, N. Carcinogenicity testing tween 7 and 12 min on chromatography of hydrolysates of of selected food additives by parenteral administration to infant Swiss mice. hepatic DNA from 1'-hydroxysafrole-treated mice is not known. Toxicol. Appi. Pharmacol., 16: 321-334, 1970. 12. Guenther, E. The Essential Oils, Vol. 1, 427 pp., 1948; Vol. 3, 777 pp., Although 1'-hydroxysafrole appeared to be bound to mouse 1949; Vol. 4, 752 pp., 1950; Vol. 5, 507 pp., 1952; Vol. 6, 481 pp., 1952. liver DNA a little more rapidly (Chart 2) than was 1'-hydroxyes New York: Van Mostranti 13. Guenther, E., and Althausen, D. The Essential Oils. Vol. 2. 852 pp. New tragole at the same dose (26), the total amounts bound and the York: Van Nostrand, 1949. shapes of the curves for the formation and loss of the DNA-, 14. Homburger, F., Kelley, T., Jr., Baker, T. R.. and Russfield, A. B. Sex effect rRNA-, and protein-bound derivatives of 1'-hydroxyestragole on hepatic pathology from a deficient diet and safrole in rats. Arch. Pathol., 73: 118-125, 1962. and 1'-hydroxysafrole were quite similar. Moreover, with both 15. Homburger, F., Kelley, T., Jr., Friedler, G., and Russfield, A. B. Toxic and compounds, there were no significant differences in the relative possible carcinogenic effects of 4-allyl-1,2-methylenedioxybenzene (safrole) in rats on deficient diets. Med. Exp., 4: 1-11, 1961. proportions of the DNA Adducts I to IV at each of the time 16. Irving, C. C., and Veazey, R. A. Isolation of deoxyribonucleic acid and points studied. Thus, although most of the adducts involving ribosomal ribonucleic acid from rat liver. Biochim. Biophys. Acta, )66: 246- A/2-guanine and /V6-adenine linkage were apparently readily 248, 1968. 17. Kadlubar, F. F., Unruh, L. E., Beland, F. A.. StrÃ¤ub.K. M., and Evans, F. E. excised from DNA in vivo, a small but significant proportion In vitro reaction of the carcinogen, N-hydroxy-2-naphthylamine. with DNA at was not removed by 20 days after treatment. Adducts involving the C-8 and N2 atoms of guanine and at the W6 atom of adenine. Carcino- reaction at N2 of guanine and derived from other types of genesis, 1: 139-150, 1980. 18. Koreeda, M., Moore, P. D., Yagi, H.. Yeh, H. J. C., and Jerina, D. M. chemical carcinogens appear to be retained in rodent liver Alkylation of polyguanylic acid at the 2-amino group and phosphate by the DNA without appreciable loss for several weeks after treatment potent mutagen (Â±)-7/3.8a-dihydroxy-9/3,10/J-epoxy-7,8,9,10-tetrahydro- (4, 19, 31, 33). The type of binding curves found for 1'- benzo[a]pyrene. J. Am. Chem. Soc.. 98. 6720-6722, 1976. 19. Kriek, E. Persistent binding of a new reaction product of the carcinogen N- hydroxysafrole (Chart 2) and 1'-hydroxyestragole (26), in hydroxy-fV-2-acetylaminofluorene with guanine in rat liver DNA in vivo. which only a minor proportion of the A/2-guanine (and N6- Cancer Res., 32: 2042-2048, 1972. 20. Leung, A. Y. Encyclopedia of Common Natural Ingredients Used in Food, adenine) residues persists in the DNA, has not been observed Drugs, and Cosmetics, 409 pp. New York: Wiley-lnterscience, 1980. with other classes of chemical carcinogens. 21. Long, E. L., Nelson, A. A., Fitzhugh, O. G., and Hansen, W. H. Liver tumors Safrole and estragÃ³le belong to a large class of ring-substi produced in rats by feeding safrole. Arch. Pathol., 75. 595-604, 1963. 22. Miller, E. C., and Miller, J. A. The metabolism of chemical carcinogens to tuted allylbenzenes and propenylbenzenes which occur in reactive electrophiles and their possible mechanisms of action in carcino- many higher plants that are used in the cosmetic and food genesis. In: C. E. Searle (ed.), ACS Monograph 173, Chemical Carcinogens, pp. 737-762. Washington, D. C.: American Chemical Society, 1976. flavoring industries (12, 13, 20). Studies are continuing in our 23. Moore, P. D., and Koreeda, M. Application of the change in partition laboratory on the biological properties and mechanism of action coefficient with pH to the structure determination of alkyl-substituted gua- of various compounds of this class. nosines. Biochem. Biophys. Res. Commun., 73. 459-464, 1976. 24. Oswald, E. O., Fishbein, L., Corbett, B. J., and Walker, M. P. Identification of tertiary aminomethylenedioxypropiophenones as urinary metabolites of safrole in the rat and guinea pig. Biochim. Biophys. Acta, 230: 237-247, ACKNOWLEDGMENTS 1971. 25. Phillips, D. H., Grover, P. L., and Sims, P. A quantitative determination of the The authors are grateful to Grant Grothman for excellent technical assistance. covalent binding of a series of polycyclic hydrocarbons to DNA in mouse skin. Int. J. Cancer, 23: 201-208, 1979. 26. Phillips, D. H., Miller, J. A., Miller, E. C., and Adams, B. Structures of the REFERENCES DNA adducts formed in mouse liver after administration of the proximate hepatocarcinogen 1'-hydroxyestragole. Cancer Res., 41: 176-186, 1981. 1. Abbott, D. D., Packman, E. W., Wagner, B. M., and Harrisson. J. W. E. 27. Sims. P. Metabolic activation of chemical carcinogens. Br. Med. Bull., 36: Chronic oral toxicity of oil of sassafras and safrol. Pharmacologist, 3: 62, 11-18, 1980. 1961. 28. Stillwell, W. G., Carman, M. J., Bell, L.. and Horning, M. G. The metabolism 2. Baird, W. M.. and Brookes, P. Isolation of the hydrocarbon-deoxyribonu- of safrole and 2',3'-epoxysafrole in the rat and guinea pig. Drug. Metab. cleoside products from the DNA of mouse embryo cells treated in cultures Dispos.. 2: 489-498, 1974. with 7-methylbenz(aJanthracene-3H. Cancer Res., 33. 2378-2385. 1973. 29. Swanson, A. B., Chambliss, D. D., Blomquist, J. C., Miller, E. C., and Miller, 3. Baird, W. M., and Dipple, A. Photosensitivity of DMA-bound 7,12-dimethyl- J. A. The mutagenicities of safrole, estragÃ³le, eugenol, frans-anethole, and benz[a]anthracene. Int. J. Cancer, 20: 427-431. 1977. some of their known or possible metabolites for Salmonella typhimurium 4. Beland, F. A., TullÃs,D. L, Kadlubar, F. F., StrÃ¤ub,K. M., and Evans, F. E. mutants. MutÃ¢t.Res., 60. 143-153, 1979. Characterization of DNA adducts of the carcinogen fV-methyl-4-aminoazo- 30. Swanson, A. B., Miller, E. C., and Miller, J. A. The side-chain epoxidation benzene in vitro and in vivo. Chem.-Biol. Interact., 3). 1-17, 1980. and hydroxylation of the hepatocarcinogens safrole and estragÃ³le and some 5. Borchert, P., Miller, J. A., Miller, E. C.. and Shires, T. K. 1'-Hydroxysafrole, related compounds by rat and mouse liver microsomes. Biochim. Biophys.

2670 CANCER RESEARCH VOL. 41

Acta. 673. 504-516, 1981. bound form of the carcinogen W-acetyl-2-aminofluorene to rat liver DNA in 31. Tarpley, W. G., Miller, J. A., and Miller, E. C. Adducts from the reaction of vivo. Chem.-Biol. Interact., 15: 149-164, 1976. W-benzoyloxy-W-methyl-4-aminoazobenzene with deoxyguanosine or DMA 34. Wislocki, P. G.. Borchert. P., Miller, J. A., and Miller, E. C. The metabolic in vitro and from hepatic DMA of mice treated with N-methyl- or N. N-dimethyl- activation of the carcinogen 1'-hydroxysafrole in vivo and in vitro and the 4-aminoazobenzene. Cancer Res., 40: 2493-2499, 1980. electrophilic reactivities of possible ultimate carcinogens. Cancer Res., 36. 32. Tomasz, M. Extreme lability of the C-8 proton: a consequence of 7-methyl- 1686-1695. 1976. ation of guanine residues in model compounds and in DNA and its analytical 35. Wislocki, P. G., Miller, E. C., Miller, J. A., McCoy, E. C., and Rosenkranz, H. application. Biochim. Biophys. Acta, 199: 18-28, 1970. S. Carcinogenic and mutagenic activities of safrole, 1'-hydroxysafrole, and 33. Westra, J. G., Kriek, E., and Hittenhausen, H. Identification of the persistently some known or possible metabolites. Cancer Res., 37: 1883-1891. 1977.

JULY 1981 2671

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 1981 American Association for Cancer Research. N2 Atom of Guanine and N6 Atom of Adenine Residues as Sites for Covalent Binding of Metabolically Activated 1 ′ -Hydroxysafrole to Mouse Liver DNA in Vivo

David H. Phillips, James A. Miller, Elizabeth C. Miller, et al.

Cancer Res 1981;41:2664-2671.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/41/7/2664

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/41/7/2664. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.