Chem. Res. Toxicol. 1991, 4, 318-323
318
An Unusual Dearomatlzed Adduct Formed by Reaction of 4’-Fluoro-4-( acetylamlno)biphenyl N-Sulfate with Deoxyadenosine Monique L. M. van de Pol1,t Vicky Venizdos,? Wilfried M. A. Niessen,s and John H. N. Meerman*ft Division of Toxicology and Division of Analytical Chemistry, Center for Bio-Pharmaceutical Sciences, Sylvius Laboratories, University of Leiden, P.O. Box 9503, 2300 RA Leiden, The Netherlands Received November 15,1990 T h e sulfate ester of the liver carcinogen N-hydroxy-4’-fluoro-4-(acetylamino)biphenyl (FAABP-N-sulfate) is believed to be a reactive intermediate in the male rat liver in vivo. After reaction of FAABP-N-sulfate with double-stranded calf thymus DNA in vitro, 30% of the adduds was identified as N-(deoxyguanosin-&yl)-4’-fluoro-4-(acetylamino)biphenyl (dG-CSFAABP) and 16% was suggested to be 3-(deoxyguanosin-N2-yl)-4’-fluoro-4-(acetylamino)biphenyl. T o investigate the identity of the remaining adducts, FAABP-N-sulfate was reacted with deoxyadenosine. Two adducts could be isolated, which were identified by ‘H NMR and mass spectrometry as 3-(deoxyadenosin-NG-yl)-4’-fluoro-4-(acetylamino)biphenyl(3-dA-NG-FAABP) and 3-(deoxyadenosin-~-yl)-4’-fluoro-4-(acetylimino)-3,4-dihydrobiphenyl(3-dA-NG-FHAIBP). An additional center of chirality is introduced at C3 (biphenyl) in the latter adduct. Therefore, 3-dA-NG-FHAIPB exists as a pair of two diastereomers with H-3 (biphenyl) in the a or j3 position. Hydrogen bonding between the proton on N8 (adenine) and the imine nitrogen or the acetylimino oxygen is suggested to stabilize 3-dA-Ne-FHAIBP and to prevent its conversion to 3-dA-NGFAABP by restoration of the aromatic system. The adduct 3-dA-NG-FHAIBP was also formed in the reaction of N-OS03H-FAABP with DNA; it accounted for 3 4 % of total covalent binding.
Introductlon Recently, we have compared various biological effects of the structurally related carcinogens N-hydroxy-24acety1amino)fluorene (N-OH-AAF),’ N-hydroxy-4’-fluoro4-(acetylamino)biphenyl (N-OH-FAABP), and Nhydroxy-4-(acetylamino)biphenyl(N-OH-AABP)in the rat liver because these compounds have widely varying hepatocarcinogenic activities (I). All three compounds were able to induce preneoplastic foci in the liver when administered to male Wistar rats after a 2/3 partial hepatectomy. They largely differed, however, in their ability to promote the outgrowth of diethylnitrosamine-initiated hepatocytes to preneoplastic foci: N-OH-AAF exhibited high promoting activity whereas N-OH-FAABP and N OH-AABP showed low and no activity respectively. A similar difference was seen in their ability to induce DNA double strand breaks as measured by the formation of micronuclei in rat liver in vivo. The clastogenicity of N-OH-AAF is most likely due to the formation of the N-acetylated DNA adduct N-(deoxyguanosin-8-y1)-2(acety1amino)fluorene(dG-C8-AAF)and is not related to the formation of the deacetylated DNA adduct N-(deoxyguanosin-8-yl)-2-aminofluorene (dG-C8-AF) (2) (dG-C8AAF and dG-CBAF were the two main adducts formed in rat liver in vivo). To investigate whether the low clastogenic potential and promotion activity of N-OHAABP and N-OH-FAABP are due to formation of only small amounts of N-acetylated DNA adducts of these compounds, DNA samples from [ring-3H]-N-OH-AABPand [rir~g-~Hl-N-OH-FAABP-treated animals were analyzed ( 3 , 4 ) . The amounts of dG-C8-AABP and dG-C8-
* To whom correspondence should be addressed. of Toxicology. Division of Analytical Chemistry.
t Division t
0893-228x/91/2704-0318$02.50/0
F M P were indeed much lower than the amount of dGC8-AAF formed after administration of N-OH-AAF. However, the formation of the deacetylated adducts N(deoxyguanosin-8-yl)-4-aminobiphenyl(dG-C%ABP)and N-(deoxyguanosin-8-yl)-4’-fluoro-4-aminobiphenyl (dGC8-FABP) was also lower compared to the formation of N- (deoxyguanosin-8-y1)-2-aminofluorene (dG-C8-AF). Furthermore, in both the AABP- and the FAABP-modified DNA samples, a large amount of radioactivity did not coelute with either of the two synthetic adducts. Gupta and Dighe (5),using 32P-postlabeling, found two additional DNA adducts (besides dG-C&AABP and dG-CS-ABP) in rat liver after administration of N-OH-AABP: one nonacetylated adduct and one acetylated adduct [presumably 3-(deoxyguanosin-N2-y1)-4-(acetylamino)bipheny1].Also, after administration of N-OH-FAABP two acetylated DNA adducts have been found one was identified as dG-C8FAABP whereas the other was suggested to be 34deoxyguanosin-N2-4’-fluoro-4-(acetylamino)biphenyl(6). Thus, most likely, various acetylated DNA adducts are formed in vivo from the two biphenyl carcinogens,the identity of which has not been established. Therefore, we characterized two FAABP-deoxyadenosine adducts formed in vitro upon reaction of the N-sulfate of N-OH-FAABP, Abbreviations: N-OH-AAF, N-hydroxy-2-(acety1amino)fluorene; N-OH-FAABP, N-hydroxy-4’-fluoro-4-(acetylamino)biphenyk N-OHAABP, N-hydroxy-4-(acetylaino)biphenyl;dG-CE-FABP, N-(deoxyguanosjn-Syl)-4’-fluoro-4-aminobiphenyl;dG-CSFAABP, N-(deoxyguanos1n-8-yl)-4’-fluoro-4-(acetylamino)bipheny1; guanin-C8-FAABP, N-(guanin-8-yl)-4’-fluoro-4-(acetylamino)bipheny1; 3-dA-M-FHAIBP, 3-(deox adenosin-N8-yl)-4’-fluoro-4-(acetylimino)-3,4-dihydrobiphenyl; 3-dA-d-FAABP, 3-(deoxyadenwin-M-yl)-4’-fluoro-4-(acetylamino)biphenyl; FAABP-N-sulfate,N-sulfate of N-OH-FAABP;TFA, trifluoroacetic acid; BTE, aqueous 5 mM [bis(2-hydroxyethyl)amino~tris(hydroxymethy1)methane and 0.1 mM EDTA, pH 7.1, buffer; dsDNA, double-stranded calf thymus DNA; -DNA, single-stranded (denatured) calf thymus DNA; FAB-MS, fast atom bombardment mass spectrometry; LC-MS, liquid chromatography-mess spectrometry.
0 1991 American Chemical Society
dAdo Adducts of 4'-Fluoro-4- (acetylamino)biphenyl
which is believed to be a reactive intermediate in vivo (3), with calf thymus DNA.
Chem. Res. Toxicol., Vol. 4, No. 3, 1991 319
MgC12was added to a final concentration of 10 mM. The DNA samples were incubated at 37 OC with DNase (0.1 mg/mg of DNA) for 6 h and with alkaline phosphatase (1.0 unit/mg of DNA) and venom phosphodiesterase (0.04 unit/mg of DNA) for 17 h. To Materials and Methods check whether the hydrolysis was complete, an aliquot of the Chemicals. dG-C8-FAABP, dG-CSFABP, and [ri~zg-~HI-N- sample was separated by reversed-phase HPLC (system C); the chromatogram was compared to that of an aqueous solution of OH-FAABP were from sources as described before (3). [ringthe four deoxyribonucleosides(each at a concentration of 4 mM). 3H]dG-C&FAABP and [ring-3H]dG-C8-FABPwere prepared in Occasionally, the incubation was continued with an additional 40% and 80% yield, respectively, as described for the unlabeled amount of alkaline phosphatase (1.0 unit/mg of DNA) and venom analogues (3). Both adducts were 90-95% pure as checked by phosphodieaterase (0.04 unit/mg of DNA) for another 6-9 h After reversed-phase HPLC and liquid scintillation counting. The N-sulfate of [n'ng-%]-N-OH-FAABP ([Ting-W]FAAsP-N-sulfate) hydrolysis, protein was precipitated with an equal volume of methanol, the samples were centrifuged,and the supernatant was was prepared similarly as described in ref 3 for the unlabeled transferred to another tube. The unlabeled synthetic deoxyanalogue. Nfl-Dimethylformamide (analytical reagent grade) guanosine adducts dG-CSFAABP and dG-C&FABP were added was from Janssen, Beerse, Belgium. TFA was purchased from to the samples to serve as UV markers. After lyophilization,the Merck, Darmstadt, FRG. Calf thymus DNA (type I), deoxysamples were dissolved in 50% (v/v) methanol/aqueous 50 mM adenosine, bacterial alkaline phosphatase (typeI - S ) , and venom potansium phosphate buffer @H 5.5) and the deoxyribonucleosides phosphodiesterase I (type VII) were obtained from Sigma were separated by reversed-phase HPLC (system A). Chemicals Co., St. Louis, MO. Pancreatic deoxyribonuclease I For TFA hydrolysis of FAABP-modified ssDNA and dsDNA, (DNase I, grade 11) was purchased from Boehringer, Mannheim, the synthetic W markers dGC8FAABP and dG-CSFABP were FRG. [81W]Deoxyadenosine(sp act. 54.9 mCi/mmol) was from added to the DNA samples. The samples were lyophilized and NEN Research Products. hydrolyzed with 1mL of TFA at 70 "C for 45 min in the presence HPLC System A for t h e Separation of DNA Adducts. A of 2% water. TFA was evaporated with a stream of nitrogen, and 120 Nucleosil5C18 column (13 X 0.39 cm) (Machery & Nagel, the residue was dissolved in 150 pL of 50% methanol/aqueous Diiren, FRG) connected to a 120 Nucleosil7C18 column (15 X 50 mM potassium phosphate buffer (pH 5.5). The nucleobases 0.39 cm) (Machery & Nagel, Diiren, FRG) was used. Elution was were separated by using HPLC system A. with methanol/aqueous 50 mM potassium phosphate buffer (pH To test the stability of synthetic 3H-labeleddG-CSFAABPand 5.5) at a flow of 1mL/min. The following gradient was used with dG-C8FABP during TFA and enzymatic hydrolysis, the adducts a linear rate of increase between each step: 5% MeOH at t = were added each to a solution of 1mg/mL unmodified calf thymus 0 min, 15% MeOH at t = 1min, 35% MeOH at t = 6 min, 45% DNA in BTE. Aliquots of these samples were hydrolyzed enMeOH at t = 12 min, 65% MeOH at t = 35 min, and 70% MeOH zymatically or with TFA as decribed in this section. at t = 40 min. Detection was at 280 nm. Reaction of [ring-3H]FAABP-N-sulfate with [S-l'CIHPLC System B for t h e Isolation of Deoxyadenosine Deoxyadenosine. A 10.2 mM solution of [ring-3H]FAABP-NAdducts. A 100 Nucleosil7C18 column (25 X 1cm) (Machery sulfate (150 pL; sp act. 7.1 mCi/mmol) in anhydrous Nfl-di& Nagel, Diiren, FRG) was used. Elution was with methanol/ methylformamide was added to 900 pL of a 4.4 mM solution of aqueous 50 mM ammonium formate buffer (pH 5.5) at a flow of [8-14C]de~xyaden~sine (sp act. 0.61 mCi/mmol) in aqueous 0.01 4 mL/min. The following gradient was used with a linear rate M potassium phosphate buffer (pH 7.4) and incubated for 2 h of increase between each step: 20% MeOH at t = 0 min, 40% at 37 OC. A control incubation was done without [8-"C]deoxyMeOH at t = 7 min, 40% MeOH at t = 20 min, and 70% MeOH adenosine. The samples were extracted three times with an equal at t = 40 min. Detection was at 280 nm. volume of diethyl ether, and an aliquot of the aqueous phase was HPLC System C for t h e Separation of Unmodified Deseparated by reversed-phase HPLC (system A). Fractions of 0.5 oxyribonucleotides. A 120 Nucleosil7C18 column (15 X 0.39 min were collected, and the radioactivity was measured by liquid cm) (Machery & Nagel, Diiren, FRG) was used. Elution was with scintillation counting in Emulsifier Safe (Packard Instruments methanol/aqueous 50 mM potassium phosphate buffer (pH 5.5) BV, Groningen, The Netherlands). using a linear gradient from 0% to 33% methanol in 33 min and For the isolation of putative deoxyadenosine adducts, a similar a flow of 1 mL/min. Detection was at 254 nm. incubation was performed now with unlabeled deoxyadenosine Reaction of [rj~g-~H]FAABP-N-sulfate with Double(200 mL, 10 mM) to which was added unlabeled FAABP-N-sulfate Stranded and Single-Stranded Calf Thymus DNA. Dou(23 mL, 17.5 mM). In the control incubation, 1 mL of 17.5 mM ble-stranded calf thymus DNA (dsDNA) was dissolved in 0.01 FAABP-N-sulfate in anhydrous DMF was added to 10 mL of M potassium phosphate buffer (pH 7.4) at a concentration of 0.5 aqueous 0.01 M potassium phosphate buffer (pH 7.4). An aliquot mg/mL. For the preparation of single-stranded DNA (ssDNA), of the samples was separated by reversed-phase HPLC (system dsDNA was denatured during 15 min at 100 "C and immediately A). Components that were only formed in the incubation with cooled on ice thereafter. [rir~g-~H]FAABP-N-sulfate (100 pL; sp deoxyadenosine and not in the control incubation were isolated act. 74 mCi/mmol) in anhydrous N,N-dimethylformamide was by preparative reversed-phase HPLC (system B). An aliquot of added to 1 mL of DNA solution and incubated for 2 h at 37 "C. the isolated fractions was checked for purity by using HPLC The concentration of [rir~g-~H]FAABP-N-sulfate was 0.5 or 1.0 system A; the remainder was lyophilized. The identity of the mM in incubations with ssDNA (resulting in a covalent binding compounds was established by 300- and 400-MHz 'H NMR and of 47 and 100 nmol/mg of DNA, respectively) and 0.2 mM in mass spectrometry (MS). The 'H NMR spectra were recorded incubations with dsDNA (resulting in a covalent binding of 4 in dimethyl sulfoxide on Bruker WM 300 and WM 400 specnmol/mg of DNA). The samples were extracted twice with 2 trometers. Mass spectral analyses were carried out on a Finnigan volumes of chloroform/isoamylalcohol/phenol(241:25) and twice MAT TSQ70 mass spectrometer (Finnigan MAT, San Jos6, CA). with 2 volumes of diethyl ether. After addition of 50 pL of a Fast atom bombardment mass spectral analysis (FAB-MS) was saturated aqueous sodium acetate solution, DNA was precipitated performed with glycerol as matrix solvent and 8-keV Argon. For with 5 volumes of ethanol (-20 "C). The DNA samples were liquid chromatography-mass spectrometry (LC-MS) the mass washed once with 70% (v/v) ethanol/H20 and once with 98% spectrometer was equipped with a Finnigan MAT thermospray (v/v) ethanol/H20, dried in vacuo, and dissolved in 5 mM Bis-Tris interface and the components were separated by using an ODS and 0.1 mM EDTA, pH 7.1 (BTE). The concentration of DNA Ultrasphere (5-pm) column (250 x 4.6 mm) that was eluted with was calculated from the absorbance at 260 nm, and the covalent methanol/aqueous 50 mM ammonium acetate (0-70% MeOH, binding was determined by liquid scintillation counting in linear gradient 0-40 min) at a flow of 1.5 mL/min. UV detection Plasmasol (Packard Instrument BV, Groningen, The Netherwas at 254 nm. lands). Enzymatic and TFA Hydrolysis. For enzymatic hydrolysis of FAABP-modified ssDNA or dsDNA dissolved in aqueous 5 mM Results [bis(2-hydroxyethyl)amino]tris(hydroxymethy1)methane and 0.1 H y d r o l y s i s of [ ~ ~ I ~ - ~ H ] ~ G - C ~ - and FAA [ringBP EDTA, pH 7.1, buffer (BTE), 7% (v/v) 1 M Tris-HC1 (pH 8.0) was added to the samples to adjust the pH of the buffer to 8.0. 3H]dG-C8-FABP during T F A and E n z y m a t i c H y -
320 Chem. Res. Toricol., Vol. 4, No. 3, 1991
0
5
10
75
20
21
30
35
van de Poll et al.
60
d5
0
5
10
13
20
25
30
35
40
43
time (mini
time (mid
Figure 1. HPLC profile of TFA-hydrolyzedasDNA modified by reaction with [ring-sH]FAABP-N-sulfate. (-) W tracing of the guanine analogues of dG-C8-FAABP(a) and dG-CB-FABP (b); the shaded bars represent the radioactivity per 5-min fraction.
Figure. 2. HPLC profile of TFA-hydrolyzed dsDNA modified by reaction with [n'ng-3H]FAABP-N-sulfate. (-) UV tracing of the guanine analogues of dG-CS-FAABP (a) and dG-CS-FABP (b); the shaded bars represent the radioactivity per Smin fraction. Peak c is suggested to be dG-M-FAABP and accounts for 16% of the radioactivity in both TFA and enzymatic hydrolysates.
drolysis of Unmodified Calf Thymus DNA. The synthetic 3H-labeleddeoxyguanosine adducts dG-C&FAABP and dG-C&FABP were each added to 250 rL of a solution of unmodified calf thymus DNA (1mg/mL) in BTE and hydrolyzed with TFA. Separation of the nucleobases after TFA hydrolysis by reversed-phase HPLC showed that, for each adduct, 90-95% of the radioactivity eluted as one single peak with a slightly longer retention time than that of the deoxyguanosine analogue. Similar recoveries were found when the synthetic adducts were incubated under the conditions of enzymatic hydrolysis. TFA and Enzymatic Hydrolysis of ssDNA and dsDNA Reacted with [ring-3H]FAABP-N-sulfate. After TFA hydrolysis of FAABP-modified ssDNA and separation of the nucleobases by reversed-phase HPLC, 77% of the radioactivity recovered from the column coeluted with N-(guanin-8-yl)-4'-fluoro-4-(acetylamino)biphenyl (guanin-CBFAABP), which is formed upon acid hydrolysis of the UV marker dG-C&FAABP. Some minor adducts were formed which accounted for the remainder of 3H activity (Figure 1). After enzymatic hydrolysis of an aliquot of the same DNA sample and separation of the deoxyribonucleosides by reversed-phase HPLC, only 50% of the 3H activity coeluted with the UV marker dG-C8FAABP. A major amount of radioactivity (35%) eluted before dG-C8-FAABPa When the remainder of the enzymatically hydrolyzed sample was lyophilized and subsequently hydrolyzed to the corresponding bases with TFA, 72% of the 3H activity recovered from the column coeluted with guanin-C&FAABP, indicating an incomplete recovery of dG-C8-FAABP in the enzymatic hydrolysis. When [ring-3H]FAABP-N-sulfate was reacted with dsDNA, several other adducts besides dG-C&FAABP were formed in considerable amounts as became evident after TFA hydrolysis (Figure 2). Enzymatic hydrolysis again seemed to be incomplete especially with regard to the recovery of dG-C8-FAABP: only 16% of the radioactivity coeluted with the UV marker dG-C8-FAABP whereas 30% coeluted with guanin-CSFAABP after TFA hydrolysis. When enzymatically hydrolyzed samples were subsequently hydrolyzed with TFA, the percentage of radioactivity coeluting with guanin-C&FAABP was again ~ 3 0 % . Reaction of FAABP-N-sulfate with Deoxyadenosine. After incubation of [8-14C]deoxyadenosine with [n'ng-3H]FAABP-N-sulfate, =95% of the 3H activity (but no "C activity) was diethyl ether extractable. An aliquot of the aqueous phase was analyzed by reversedphaae HPLC (system A), and fractions were collected every 5 min. UV peaks were seen in the HPLC chromatogram which were not present in that of the control incubation (without [ 8-'4C]deoxyadenosine). Comparison of the
amount of 3H activity with the amount 14Cactivity coeluting with peaks IA/B, 11, and I11 eluting at 20.5/20.8, 30.2, and 32.4 min, respectively, indicated a 1:l molar ratio of a biphenyl moiety and deoxyadenosine. These peaks contained ==7.2%,0.9%, and 3.6% of the 3H activity injected onto the column. In the larger scale incubation with unlabeled compounds, similar peaks were observed after HPLC (except for peak I1 which was not found). Peaks IA and IB were collected in one fraction because they could not be separated adequately. The UV spectrum of IA/B (aqueous solution, pH 7.0) showed maxima at 263 and 208 nm and a minimum at 239 nm; the spectrum of I11 showed maxima at 267 and 206 nm and a minimum at 240 nm. Spectra taken at pH 1.0 or 10.0 were very similar: minima and maxima differed by not more than 1-2 nm, except for the spectrum of IA/B at pH 10.0 which showed an additional shoulder at 300-310 nm while the absorbtion at 263 nm was lower. The 'H NMR spectra of IA/B and I11 showed signals (Table I) for the deoxyadenosine protons at chemical shifts comparable to those reported before for l-(deoxyadenosin-IP-yl)-2-naphthylamine(7) and 3-(deoxyadenosin-IP-y1)-N-methylaminoazobenzene (8). No signal for an exocyclic 6-amino group was present at 6 7.35 ppm; instead, a signal was seen from one N6-H proton at 6 9.89 (or 9.17) ppm for adduct I11 and at 6 9.52 ppm for adducts IA/B. This would indicate a substitution a t N6 of deoxyadenosine in both IA/B and 111. Interestingly, however, the N6-H proton in IA/B was not readily exchangeable with D20 whereas it was in compound 111. Also, there is a difference in chemical shifts between IA/B and I11 especially with regard to the adenine protons which for IA/B appear ~0.3-0.4ppm upfield from the corresponding protons of adduct 111. For the FAABP moiety of adduct I11 (Figure 3A), a signal was found for the N-acetyl group at 6 2.09 ppm and for the 4-amide proton at 6 9.17 (or 9.89) ppm (exchangeable with D20). The aromatic protons were assigned with the aid of homonuclear decoupling, and chemical shifts and coupling constants are presented in Table I. The data are compatible with a 4'-fluoro-4(acety1amino)biphenyl-deoxyadenosineadduct in which the FAABP moeity is substituted in the ring ortho to the amido group and deoxyadenosine is substituted at the exocyclic NH2. The identity of I11 as an FAABP-deoxyadenosine adduct was confirmed by FAB-MS: the spectrum showed an [M + H]+ at m / z 479. Thus, adduct I11 was identified as 3-(deoxyadenosin-IP-y1]-4'-fluoro-4(acetylamino)biphenyl (3-dA-Ns-FAABP) (Figure 5). A signal was found for the N-acetyl group of adduct IA/B (Figure 3B) at the same chemical shift as seen for
Chem. Res. Toxicol., Vol. 4, No. 3, 1991 321
dAdo Adducts of 4‘-Fluoro-4-(acety1amino)biphenyl Table I. *€I NMR Data of 34Deoxyadenoclin-N‘-yl)-4’-fluoro-4-( acetylimino)-3,4-dihydrobiphenyl (3-dA-N6-FHAIBP)and 3 4Deoxyadenoein-N“-yl)-4’-fluoro-4-(acetylamino) biphenyl (B-dA-N‘-FAABP)’ Deoxyadenosine Protons multiplicity chemical shift: ppm (no. of 3-dA-M- 3-dA-Massignment protons) FHAIBP FAABP H-2’b or H-2’a (dR) m (1) 2.26 2.30 H-2’a or H-2’b (dR) m (1) 2.51 2.16 H-5’a,b (dR) m (2) 3.56 3.59 H-4’ (dR) m (1) 3.85 3.89 H-3’ (dR) m (1) 4.31 4.43 OH-5’ (dR) m (1) 5.W 5.14e OH-3’ (dR) 5.28* 5.34. H-1’ (dR) t (1) (’) 6.24 6.40 H-8 or H-2 (Ade) 1.92 8.30 H-2 or H-8 (Ade) s (1) 8.21 8.50 s (1) 9.52 9.89. N6-H (Ade) (or 9.17)’ 4’-Fluoro-4-(acetylamino)biphenylProtons of 3-dA-M-FHAIBP multiplicity chemical coupling assign(no. of ment Drotons) shift! DDm constant. Hz s (3) 2.00 CH3 4.94014.949 J3,2 5.9 H-3 dld (Ud H-5 d (1) 5.84 JKa 10.0 H-6 dd/dd (Ud 5.99616.000 J;:; = 10.0; 56.2 = 1.1 m (1) H-2 6.65 Jz,& J2,3 = 5.8 dd (2) H-3’,5’ 1.20 J3p,r,J5,,69 = 8.9; J3jp, JST = 8.9 H-2‘,6’ ddldd (2Id 1.405/1.410 Jr,s,,Ja.,s, = 8.9; J r p , J6,,p = 8.9 4’-Fluoro-4-(acetylamino) biphenyl Protons of 3-dA-M-FAABP multiplicity chemical assign(no. of shift,b coupling constant, ment protons) ppm Hz s (3) 2.09 CH3 1.28 J3zvr,J5#,6~ 8.9; H-3’,5’ dd (2) J8.p J5tz = 8.9 H-6 dd (1) 1.46 J6,5 = 8.3; J6,2 2.1 H-5 d (1) 1.51 J5s = 8.3 H-2’,6’ dd (2) 1.68 Jy,3,, J6,,5, = 8.8; J y p , J6,j = 5.5 H-2 d (1) 8.03 J2,e = 2.1 4-NH 8 (1) 9.17. ‘The lH NMR spectrum of 3-dA-NB-FHAIBP was recorded at 400 MHz and that of 3-dA-M-FAABP at 300 MHz; Ade, adenine; dR, deoxyribose; s, singelt; d, doublet; dd, double doublet; t, triplet; m, multiplet. bChemical shifts of protons (in ppm) are relative to those of tetramethylsilane. CNotpossible to quantify because of insufficient resolution. dThe signals for H-2’,6’, H-6, and H-3 are actually composed of two double doublets, two double doublets, and two doublets, respectively, with a small difference in chemical shift. e Protons exchangeable with D20.
adduct 111; a signal for the amide proton near 6 9.17 (or 9.89) ppm, however, was absent. A two-dimensional Jcorrelated (COSY)spectrum (Figure 4) was taken, and together with homonuclear decoupling experiments, this was used for the assignment of the remaining FAABP protons (Table I). Besides signals for the aromatic proton H-3’,5’ and H-2’,6’ between d 7.0 ppm and 6 7.5 ppm, signals for four nonaromatic protons appeared a t lower chemical shifts. On the basis of homonuclear decoupling data, these were assigned to H-2, H-3, H-5, and H-6. H-6 showed a strong coupling with H-5 and a weak coupling with H-2, and H-2 showed a weak coupling with H-6 and a stronger coupling with H-3. Due to the introduction of another substituent, C3 becomes a center of chirality. Therefore, adducts IA and IB might be diastereomers with H-3 in either the (Y or B position. This would also explain the almost identical chromatographic properties of the two
A
9 5
8 5
S O
1 0
I 5
PW
7 0
I 5
6 0
I 5
B O
Figure 3. ‘H NMR spectra of the FAABP protons of 3-dANs-FHAIBP (adduct IA/B) (A) and 3-dA-Ns-FAABP(adduct III) (B).Peaks x and y in panel B are signals of an unknown impurity. I
5
3
Figure 4. Two-dimensional &correlated (COSY) spectrum of 3-dA-I@-FHAIBP (adduct IA/B).
van d e Poll et al.
322 Chem. Res. Toxicol., Vol. 4, No. 3, 1991 F
\ /
\ /
H--+3 T
NH
J-dA-N'-FAABP
'
F
T
&
H
3
NH
3-dA-N'-FHAIW
Figure 5. Structures of 3-dA-M-FMP and 3-dA-P-FHAIBP.
adducts. The presence of two diastereomers is substantiated by the fact that the signal for H-3 appears as two similar sets of signals (two doublets) with a small difference in chemical shift (0.009 ppm). Similarly, the signals of H-6 and H-2',6' appear as two sets of signals. These data, together with the absence of a signal for the amide proton, are compatible with an FAABP-deoxyadenosine adduct in which the biphenyl moiety is ortho substituted with formation of an imine and deoxyadenosine is substituted at the N6 position. The identity of IA/B as an FAABPsubstituted deoxyadenosine adduct was confirmed by FAB-MS: the spectrum showed an [M + H]+ at m / z 479. No important differences were found with the spectrum of 3-dA-NG-FAABP. The sample was also analyzed by LC-MS. The two diastereomers could not be separated, but the mass spectrum of the single peak showed a fragment at m / z (relative intensity) 363 (18) corresponding to an FAABP-deoxyadenosine adduct which has lost the deoxyribose moiety. Fragments corresponding to the FAABP moiety and adenine, respectively, were seen at m / z 230 (32) and m / z 136 (100). Thus, the identity of IA/B was established as 3-(deoxyadenosin-NG-yl)-4'-fluoro-4-( acetylimino)-3,4-dihydrobiphenyl (3-dA-Ne-FHAIBP) [systematic name: 2-(acetylimino)-1-(2'-deoxyadenosin-NG-yl)(4-fluorophenyl)cyclohexa-3,5-diene; Figure 51. Depurination of 3-dA-Ns-FAABP and 3-dA-NG-FHAIBP in 0.1 M HCl at 37 "C yields compounds with similar UV spectra as the parent adducts. HPLC analysis showed that a single product was formed from each adduct (no adenine or deoxyadenosine was detected), and no conversion of 3-dA-Ne-FAABP into 3-dA-Ne-FHAIBP or vice versa was detected. The retention times of the presumed adenine analogues were identical with those of the adenine analogues obtained after TFA hydrolysis (see below). Formation of FAABP-Deoxyadenosine Adducts with DNA in Vitro. We have investigated if FAABPdeoxyadenosine adducts are formed upon reaction of [ring-3H]FAABP-N-sulfate with dsDNA. Therefore, 3dA-Ne-FAABP and 3-dA-NG-FHAIPB were first hydrolyzed with TFA to their corresponding adenine analogues and analyzed by reversed-phase HPLC (system A). The chromatogram of the adenine analogue of 3-dA-Ns-FAABP showed a single UV peak with a retention time close to that of the deoxyadenosine adduct and was, therefore, assumed to be stable during TFA hydrolysis. In the chromatogram of 3-dA-Ne-FHAIBP two UV peaks (x and y) appeared after TFA hydrolysis (Figure 6). To identify these compounds, the hydrolyzed sample was analyzed by LC-MS. In the reconstructed mass chromatogram several ions could be detected. The W peaks x and y corresponded with ions having an [M + H]+ at m/z 322 and 363, respectively. No rationale for the ions observed with x has been found yet. The mass spectrum of UV peak y shows ions at m / z (relative intensity) 363 (1001, 230 (18), and 136 (56), cor-
0
5
$0
75
20
25
33
35
40
a5
time
(md
Figure 6. HPLC profile after TFA hydrolysis of FAABP-modified dsDNA. Synthetic deoxyguanosine and deoxyadenosine adducts were added to the sample before hydrolysis to serve as UV markers. (-) UV tracing showing the guanine analogues of dG-C8-FAABP(a) and dG-C8-FABP(b) and the adenine analogues of 3-dA-IP-FHAIBP(y) and 3-dA-IP-FAABP(2). (x = unknown).
responding, respectively, to the FAABP-adenine adduct, the FAABP fragment, and adenine, and it was therefore identified as the adenine analogue of 3-dA-NG-FHAIBP. The unlabeled deoxyadenosine adducts 3-dA-Ns-FAABP and 3-dA-NG-FHAIBP and the UV markers of dG-CSFAABP and dG-FABP were added to FAABP-modified dsDNA obtained by reaction of FAABP-N-sulfate with dsDNA. This was lyophilized and then hydrolyzed with TFA. The nucleobases were separated by reversed-phase HPLC (system A), and fractions of 0.5 min were collected. A major amount of activity coeluted again with guaninC8-FAABP; no radioactivity coeluted with the adenine analogue of 3-dA-NG-FAABP whereas =3-6% of the radioactivity recovered from the column coeluted with the adenine analogue of 3-dA-NG-FHAIBP (Figure 6).
Discussion After TFA hydrolysis of FAABP-modified dsDNA (obwith tained by reaction of [ring-3H]FAABP-N-sulfate dsDNA), a large amount of radioactivity (70%) did not coelute with the guanine analogues of dG-CBFAABP or dG-C&FABP. This is unlikely the result of breakdown of these adducts during TFA hydrolysis because the synthetic 3H-labeled guanine adducts were stable under these conditions. In addition, enzymatic hydrolysis resulted in an even lower yield of dG-C&FAABP. This could be improved by subsequent TFA hydrolysis of an already enzymatically hydrolyzed sample, but also, in this case, 70% of the radioactivity could not be assigned to dG-CB FAABP or dG-Cg-FABP. This indicates that other, as yet unidentified, adducts are formed. One of the major adducts (c, Figure 2) might well be 3-(deoxyguanosin-Nyl)-4'-fluoro-4-(acetylamino)biphenyl (dG-N-FAABP),an adduct suggested to be formed in vivo in the rat liver after administration of N-OH-FAAEiP (6). Its identity has never been confirmed, but the fact that it is only formed in the reaction of FAABP-N-sulfate with dsDNA indicates that this adduct might indeed be dG-N-FAABP because the analogous aminofluorene adduct 3-(deoxyguanosin-Nyl)-2-(acetylamino)fluorene (dG-N-AAF) is also only formed in dsDNA (9). Besides deoxyguanosine adducts, deoxyadenosine adducts have been frequently found in vivo after exposure to arylamine carcinogens (10). Therefore, we synthesized FAABP-deoxyadenosine adducts by reaction of deoxyadenosine with FAABP-Nsulfate, which is believed to be a reactive intermediate in the metabolism of N-OH-FAABP in vivo (3). Two deoxyadenosine adducts were isolated from the incubation
Chem. Res. Toxicol., Vol. 4, No. 3, 1991 323
dAdo Adducts of 4‘-Fluoro-4- (acety1amino)biphenyl
mixture: one adduct was identified by 300-MHz ‘H NMR and mass spectrometry as 3-(deoxyadenosin-Ns-y1)-4’fluoro-4-(acetylamino)bipheny1(3-dA-Ns-FAABP) and the other as 3-(deoxyadenosin-IP-y1)-4‘-fluoro-3-hydro-4(acety1imino)biphenyl (3-dA-Ns-FHAIBP). With the dhruption of the aromatic system in 3-dA-Ns-FHAIPB an additional center of chirality is introduced a t C3. Therefore, the adduct might exist as a mixture of two diastereomers with H-3 (biphenyl) in the a or fl position. The presence of two identical sets of signals for the biphenyl H-3 proton (and the H-6 and H-2’,6’ and H-2’,6’ protons) in the ‘H NM’R spectrum is an indication for this and explains why two compounds are found which cannot be separated completely by preparative reversed-phase HPLC because of their almost identical chromatographic behavior. How these two very similar adducts (3-dA-Ns-FAABP and 3-dA-Ns-FHAIBP)are formed and can exist without conversion of one into the other is uncertain. The fact that 3-dA-Ns-FHAIBP does not rearrange into 3-dA-PFAABP suggests that the two adducts are not formed in the same way. No data in support of particular reaction mechanisms are available yet. The existence of 3-dANs-FHAIBP is puzzling because of the disruption of the aromatic system. The reason why this structure is favored and the aromatic system is not restored might be a stabilizing effect of a hydrogen bond between the proton on N6 and the imine nitrogen or the acetylimino oxygen. This would also explain why the proton on N6 only slowly exchanges with D20. Analysis of dsDNA reacted with FAABP-N-sulfate in vitro showed that, of the two FAABP-deoxyadenosine adducts, only 3-dA-Ns-FHAIBP was formed. Why the other deoxyadenosine adduct (3-dA-Ns-FAABP)is not formed is not known. It is also not yet known if this adduct contributes to the mutagenic and/or clastogenic effects of N-OH-FAABP. Acknowledgment. This study was supported by a grant of the Dutch Cancer Society (Koningin Wilhelmina Fonds), Project IKW 86-94. We thank Dr. G. J. Westra, Dr. B. Zomer, and Dr. M. M. S. D. Marques for their advice on the interpretation of the ‘H NMR spectra.
Registry No. FAABP-N-sulfate, 132884-69-6; dGC&FAABP, 67764-18-5; 3-dG-iP-FAABP, 125464-41-7; 3-dA-NB-FAABP) 132884-70-9; 3-dA-NB-FHAIBP, 132884-71-0; deoxyadenosine, 958-09 8.
-
References (1) van de Poll,M. L. M., van der Hulst, D. A. M., Tates, A. D., and
Meerman, J. H. N. (1990) Correlation between clastogenicityand promotion activity in liver carcinogenesis by N-hydroxy-2acetylaminofluorene,N-hydroxy-4’-fluoro-4-acetylaminobiphenyl and N-hydroxy-4-acetylaminobiphenyl. Carcinogenesis 11, 333-339. (2) van de Poll, M. L. M., van der Hulst, D. A. M., Tates, A. D., Mulder, G. J., and Meerman, J. H. N. (1989) The role of specific DNA adducts in the induction of micronuclei by N-hydroxy-2acetylaminofluorene in rat liver in uiuo. Carcinogenesis 10, 717-722. (3) van de Poll,M. L. M., Tijdens, R. B., Vondracek, P., Bruins, A. P., Meijer, D. K. F., and Meerman, J. H. N. (1989) The role of sulfation in the metabolic activation of N-hydroxy-4‘-fluoro-4acetylaminobiphenyl. Carcinogenesis 10, 2285-2291. (4) van de Poll, M. L. M., Venizelos, V., and Meerman, J. H. N. (1990) Sulfation-dependent formation of N-acetylated and deacetylated DNA adducts of N-hydroxy-4-acetylaminobiphenyl in male rat liver in vivo and in isolated hepatocytes. Carcinogenesis (in press). (5) Gupta, R. C., and Dighe, N. R. (1984) Formation and removal of DNA adducts in rat liver treated with N-hydroxy derivatives of 2-acetylaminofluorene, 4-acetylaminobiphenyl and 2-acetylaminophenanthrene. Carcinogenesis 5,343-349. (6) Kriek, E., and Hengeveld, G. M. (1978) Reaction products of the carcinogenN-hydroxy-4-acetylamino-4’-fluorobiphenyl with DNA in liver and kidney of the rat. Chem.-Biol. Interact. 21,179-201. (7) Kadlubar, F. F., Unruh, L. E., Beland, F. A., Straub, K. M., and Evans, F. E. (1980) In uitro reaction of the carcinogen, Nhydroxy-2-naphthylamine,with DNA at the C8 and N2 atoms of guanine and at the N6 atom of adenine. Carcinogenesis 1, 139-150. (8) Tullis, D. L., Straub, K. M., and Kadlubar, F. F. (1981) A comparison of the carcinogen-DNA adducts formed in rat liver in uiuo after administration of single or multiple doses of N-methyl-4aminoazobenzene. Chem.-Biol. Interact. 38, 15-27. (9) Westra, J. G., Kriek, E., and Hittenhausen, H. (1976) Identification of the persistently bound form of the carcinogen Nacetyl-2-aminofluorene to rat liver DNA in uitro. Chem.-Biol. Interact. 15, 149-164. (10) Beland, F. A., and Kadlubar, F. F. (1985) Formation and persistence of arylamine DNA adducts in uiuo. Environ. Health Perspect. 62, 19-30
