Synthesis, Characterization, and Quantitation of a 4-Aminobiphenyl

Jingshu Guo, Peter W. Villalta, Christopher J. Weight, Radha Bonala, Francis .... János Minárovits,, Miriam C. Poirier,, Elise D. Bowman,, Ainsley W...
0 downloads 0 Views 162KB Size
Download PDF
68

Chem. Res. Toxicol. 1999, 12, 68-77

Synthesis, Characterization, and Quantitation of a 4-Aminobiphenyl-DNA Adduct Standard Frederick A. Beland,*,† Daniel R. Doerge,† Mona I. Churchwell,† Miriam C. Poirier,‡ Bernadette Schoket,§ and M. Matilde Marques| Division of Biochemical Toxicology, National Center for Toxicological Research, Jefferson, Arkansas 72079, Laboratory of Cellular Carcinogenesis and Tumor Promotion, National Cancer Institute, Bethesda, Maryland 20892, Department of Biochemistry, National Institute of Environmental Health, Budapest, H-1097 Hungary, and Centro de Quı´mica Estrutural, Complexo I, Instituto Superior Te´ cnico, Av. Rovisco Pais, P-1049-001 Lisboa, Portugal Received July 22, 1998 32P-Postlabeling is a powerful technique for the detection of DNA adducts; however, quantitation of DNA adducts by this method can result in errors due to differences in hydrolysis and labeling efficiencies between adducted and normal nucleotides. We have synthesized a DNA sample modified with 4-aminobiphenyl to serve as a quantitation standard for 32Ppostlabeling and other DNA adduct detection methodologies. [2,2′-3H]-N-Hydroxy-4-aminobiphenyl was reacted with calf thymus DNA at pH 5 to give 62 ( 0.8 adducts/108 nucleotides (mean ( SD) on the basis of 3H content. HPLC analyses following enzymatic hydrolysis to nucleosides indicated one major adduct, N-(deoxyguanosin-8-yl)-4-aminobiphenyl (dG-C8-4ABP). The adduct identity was confirmed by HPLC/electrospray ionization mass spectrometry, which indicated a modification level of 19 ( 1.7 dG-C8-4-ABP/108 nucleotides. 32P-Postlabeling analysis gave a value of 0.84 dG-C8-4-ABP/108 nucleotides, while a dissociation-enhanced lanthanide fluoroimmunoassay (DELFIA) indicated levels of 82 ( 26 and 63 ( 20 dG-C8-4ABP/108 nucleotides after enzymatic hydrolysis to nucleotides and nucleosides, respectively. The utility of the DNA adduct standard was determined by assessing the level of dG-C8-4ABP in liver DNA from mice treated with [2,2′-3H]-4-aminobiphenyl. 32P-Postlabeling analyses, based upon measuring the extent of the 32P incorporation, underestimated the levels of dGC8-4-ABP, while DELFIA, using a G-C8-4-ABP quantitation standard, overestimated the adduct levels. The adduct levels determined by HPLC/electrospray ionization mass spectrometry best reflected those obtained from 3H incorporation. When the 32P-postlabeling analyses and the DELFIA were conducted using the DNA modified in vitro with dG-C8-4-ABP as a quantitation standard, accurate estimations of the extent of in vivo formation of dG-C8-4-ABP were obtained.

Introduction 32P-Postlabeling analysis, an analytical technique developed by Randerath, Reddy, and Gupta (1, 2), is an extremely powerful method for the detection of DNA adducts from many classes of chemical carcinogens, including alkylating agents, polycyclic aromatic hydrocarbons, nitropolycyclic aromatic hydrocarbons, aromatic amines, heterocyclic aromatic amines, and mycotoxins (3). In its simplest form, DNA is enzymatically hydrolyzed to 3′-nucleotides, which are converted to 5′-32Plabeled 3′,5′-bisphosphates using [γ-32P]ATP and polynucleotide kinase. The 32P-labeled DNA adducts are then usually separated by TLC and detected by autoradiography. Since the inception of 32P-postlabeling analyses, two enhancement methods, based upon nuclease P1 digestion (4) and n-butanol extraction (5), have been introduced

* To whom correspondence should be addressed: HFT-110, NCTR, Jefferson, AR 72079. Telephone: (870) 543-7205. Fax: (870) 543-7136. E-mail: [email protected]. † National Center for Toxicological Research. ‡ National Cancer Institute. § National Institute of Environmental Health. | Instituto Superior Te ´ cnico.

that significantly improve the sensitivity of the procedure. The nuclease P1 variation (4) takes advantage of the resistance of certain 3′-nucleotide DNA adducts to the phosphatase activity of nuclease P1. As a result, nonmodified 3′-nucleotides that would be labeled by polynucleotide kinase are preferentially destroyed. In the n-butanol version (5), hydrophobic 3′-nucleotide DNA adducts are preferentially extracted into n-butanol, thus decreasing the concentration of nonmodified 3′-nucleotides that would become labeled. Due to the high specific activity of [γ-32P]ATP, 32P-postlabeling analyses have exquisite sensitivity, being able to detect one adduct in 1010 normal nucleotides (3). This, coupled with a broad range of applicability and the fact that only microgram amounts of DNA are required, makes 32P-postlabeling a technique widely used for monitoring human exposure to chemical carcinogens. The quantitation of DNA adducts detected by 32Ppostlabeling analysis is typically accomplished by relative adduct labeling (RAL1), where RAL is cpm in adducts/ cpm in total nucleotides × 1/dilution factor (3, 6). If it is assumed that 1 µg DNA ) 3 × 106 fmol of nucleotides, RAL can be converted into an absolute level of binding (e.g., fmol of adduct per microgram of DNA; 5).

10.1021/tx980172y CCC: $18.00 © 1999 American Chemical Society Published on Web 12/23/1998

4-Aminobiphenyl-DNA Adduct Standard

A number of assumptions are inherent when quantifying DNA adduct levels by RAL, the foremost being that adducted and normal nucleotides are converted to 3′nucleotides and 3′,5′-bisphosphates with equal efficiency. It has been found, however, that certain aromatic or bulky adducts are better substrates than normal nucleotides for polynucleotide kinase and thus become preferentially labeled (7). In contrast, other investigators have found certain DNA adducts to be labeled at lower than expected efficiencies (8). In addition, when using the nuclease P1 enrichment technique, it is assumed that the adducts are completely resistant to the 3′-phosphatase activity of nuclease P1; likewise, the n-butanol extraction version presupposes that the adducts are extracted with 100% efficiency. Since these assumptions may not be correct, erroneous adduct levels can result. For example, as part of an interlaboratory trial of 32Ppostlabeling methodologies, DNA adduct levels were assessed in mice treated with radiolabeled benzo[a]pyrene or 2-acetylaminofluorene (9). With both carcinogens, the values determined by RAL were 4-8-fold lower than those determined by measuring 3H incorporation. One approach that should allow better quantitation of DNA adduct levels by 32P-postlabeling and other DNA adduct detection methodologies is to use DNA standards covalently modified with carcinogens of interest at levels commonly observed in vivo. In this paper, we describe the preparation, characterization, and quantitation of a DNA adduct standard modified with the human urinary bladder carcinogen 4-aminobiphenyl at a level similar to that seen in human samples.

Materials and Methods Caution: Nitroarenes, N-arylamines, N-hydroxyarylamines, N-acetoxyarylamines, and N-acetoxyarylamides are potentially carcinogenic. They should be handled with protective clothing, in a well-ventilated fumehood. Exposure to 32P should be kept as minimal as possible, by working in a confined laboratory area with protective clothing, plexiglass shielding, Geiger counters, and body dosimeters. Waste materials must be discarded according to appropriate safety procedures. Chemicals. Calf thymus DNA was purchased from Sigma Chemical Co. (St. Louis, MO) and further purified by treatment with RNases A and T1 and proteinase K as described by Gupta (10). [2,2′-3H]-4-Nitrobiphenyl was purchased from Chemsyn Science Laboratories (Lenexa, KS) and purified by column chromatography on silica by eluting with chloroform. Its specific activity was determined to be 63 ( 2.1 mCi/mmol (mean ( SD; n ) 8) by UV spectroscopy and scintillation counting. The [2,2′3H]-4-nitrobiphenyl was converted to [2,2′-3H]-N-hydroxy-4aminobiphenyl by the method of Westra (11). [2,2′-3H]-4Aminobiphenyl was obtained by the reduction of [2,2′-3H]-4nitrobiphenyl with NH4Cl and Zn. Specifically, 21 mg of [2,2′3H]-4-nitrobiphenyl, which was suspended in 1 mL of 80% ethanol containing 45 mg of NH4Cl, was reacted with 70 mg of Zn for 3 h at 50 °C. The mixture was diluted with water and extracted with diethyl ether, and the diethyl ether was evaporated. The residue was dissolved in chloroform and purified by column chromatography on silica by eluting with chloroform. The [2,2′-3H]-4-aminobiphenyl had a specific activity of 51 ( 1 Abbreviations: Bis-Tris, bis(2-hydroxyethyl)iminotris(hydroxymethyl)methane; DELFIA, dissociation-enhanced lanthanide fluoroimmunoassay; dG-C8-4-ABP, N-(deoxyguanosin-8-yl)-4-aminobiphenyl; dG-C8-4-ABP-d9, N-(deoxyguanosin-8-yl)-4-aminobiphenyl-d9; FCS, fetal calf serum; G-C8-4-ABP, N-(guanosin-8-yl)-4-aminobiphenyl; HSA, human serum albumin; PBS/Tween, phosphate-buffered saline containing 0.05% Tween-20; RAL, relative adduct labeling.

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 69 0.04 mCi/mmol (mean ( SD; n ) 2), as determined by UV spectroscopy and scintillation counting. N-(Deoxyguanosin-8-yl)-4-aminobiphenyl (dG-C8-4-ABP) and N-(guanosin-8-yl)-4-aminobiphenyl (G-C8-4-ABP) were synthesized by reacting N-acetoxy-N-(trifluoroacetyl)-4-aminobiphenyl with deoxyguanosine and guanosine, respectively (12, 13). N-(Deoxyguanosin-8-yl)-4-aminobiphenyl-d9 (dG-C8-4-ABP-d9) was similarly prepared substituting N-acetoxy-4-aminobiphenyld9 for N-acetoxy-N-(trifluoroacetyl)-4-aminobiphenyl. The N-acetoxy-4-aminobiphenyl-d9 was prepared from biphenyl-d10 (Aldrich Chemical Co., Milwaukee, WI) by nitration with ammonium nitrate in the presence of trifluoroacetic anhydride (14), followed by reduction to N-hydroxy-4-aminobiphenyl-d9 (11), and Oacetylation with acetyl cyanide (15). The adducts were purified by chromatography on Sephadex LH-20 (Pharmacia/PL Biochemicals, Piscataway, NJ) using a 20 to 80% step gradient of aqueous methanol. The adducts eluted with 70% methanol and were quantified on the basis of their published molar extinction coefficients (12, 16). The 3′-nucleotide of dG-C8-4-ABP was prepared by reacting N-acetoxy-4-aminobiphenyl with deoxyguanosine 3′-phosphate. The adduct was purified on a Waters C18 Sep-Pak cartridge (Waters Associates, Milford, MA) by eluting with a 20 to 80% step gradient of acetonitrile in 100 mM ammonium acetate and 10 mM ammonium phosphate buffer (pH 5.7; 15). In Vitro Reaction of Calf Thymus DNA with N-Hydroxy-4-aminobiphenyl. To prepare a DNA standard modified with 4-aminobiphenyl, calf thymus DNA (360 mg) in 500 mL of 10 mM sodium citrate buffer (pH 5) was extensively purged with argon and then treated overnight with 3.8 µg of [2,2′-3H]-Nhydroxy-4-aminobiphenyl (63 mCi/mmol, molar ratio of 1.8 × 10-5:1 N-hydroxy-4-aminobiphenyl:DNA nucleotides) that had been dissolved in 50 mL of argon-purged ethanol. The mixture was extracted three times with distilled phenol that had been saturated with a 50 mM bis(2-hydroxyethyl)iminotris(hydroxymethyl)methane (Bis-Tris), 1 mM EDTA buffer (pH 7.1) and three times with water-saturated distilled n-butanol, and the DNA was precipitated by the addition of sodium chloride and ethanol. The DNA was collected by centrifugation, washed with 70% ethanol, and redissolved in a 5 mM Bis-Tris, 0.1 mM EDTA buffer (pH 7.1) at an approximate concentration of 1 mg/ mL. The DNA concentration was determined by UV spectrometry by assuming 48 µg ) 1.0 absorbance unit at 260 nm. The extent of binding was quantified with a Packard Tri-Carb model 1600TR scintillation counter (Packard Instrument Co., Meriden, CT), using Ultima Gold (Packard) as the scintillation fluid. Samples (1 mg) were counted in triplicate after treatment with DNase I (0.1 mg/mg of DNA) for 1 h at 37 °C. Similar incubations were conducted to obtain higher levels of modification using 3.6 mg of DNA in 5 mL of 10 mM sodium citrate buffer (pH 5) and 380 ng, 3.8 µg, and 38 µg of [2,2′-3H]N-hydroxy-4-aminobiphenyl in either 500 or 700 µL of ethanol. This gave molar ratios of 1.8 × 10-4:1, 1.8 × 10-3:1, and 1.8 × 10-2:1 N-hydroxy-4-aminobiphenyl:DNA nucleotides, respectively. A control incubation was also conducted with 35 mg of DNA in 48 mL of 10 mM sodium citrate buffer (pH 5) and 4.8 mL of ethanol. In Vivo Modification of DNA with 4-Aminobiphenyl. Eight-week-old male B6C3F1 mice (four per group; obtained from the breeding colony at the National Center for Toxicological Research) were treated intraperitoneally with 0, 0.1, or 1.0 mg of [2,2′-3H]-4-aminobiphenyl (51 mCi/mmol) in 100 µL of tricaprylin (Sigma). After 24 h, the mice were killed by exposure to CO2, livers were pooled by group, nuclei were prepared by the method of Basler et al. (17), and DNA was isolated from the nuclei with slight modifications of the method described by Beland et al. (18). The average yield of DNA was 1.3 mg/g of liver, with an average A260/A280 of 1.81. Approximately 1 mg of DNA was counted for radioactivity after treatment with DNase I as described above. Enzymatic Hydrolysis and HPLC Analysis of DNA Reacted in Vitro with N-Hydroxy-4-aminobiphenyl. DNA

70 Chem. Res. Toxicol., Vol. 12, No. 1, 1999 reacted in vitro with [2,2′-3H]-N-hydroxy-4-aminobiphenyl was hydrolyzed to nucleosides by treatment with DNase I, followed by treatment with alkaline phosphatase and snake venom phosphodiesterase as described by Heflich et al. (19). Nonradioactive dG-C8-ABP was added to the samples to serve as a UV marker, and the samples were analyzed by reversed-phase HPLC using a µBondapak C18 column (0.39 cm × 30 cm; Waters Associates) with an HPLC system consisting of two Waters model 510 pumps, a Rheodyne model 7125 injector (Rheodyne, Cotati, CA), and a Waters model 660 automated gradient controller. The peaks were monitored at 280 nm with a HewlettPackard 1050 diode array spectrophotometric detector (HewlettPackard, Wilmington, DE). Samples were eluted with a 30 min nonlinear gradient (Waters curve #2) of 20 to 56% methanol, followed by a 3 min linear gradient to 100% methanol and a 7 min isocratic elution at 100% methanol. The flow rate was 2 mL/min. One minute fractions were collected for measurement of radioactivity. When large quantities of DNA (i.e., 2 or 20 mg) were hydrolyzed, the nucleoside adducts were extracted three times with water-saturated n-butanol. The n-butanol extracts were combined, back-extracted twice with n-butanol-saturated water, and evaporated, and the residue was redissolved in methanol for HPLC analysis. 32P-Postlabeling Analysis. 32P-Postlabeling analyses were conducted with the n-butanol enrichment procedure of Gupta (5). Samples were applied to 10 cm × 10 cm polyethyleneiminecellulose TLC plates manufactured by Macherey-Nagel (Alltech Associates, Inc., Deerfield, IL). Adducts were resolved by multidirectional chromatography using the following solvents: D1 and D4, 900 mM sodium phosphate (pH 6.8); D2, 3.6 M lithium formate and 8.5 M urea (pH 3.5); and D3, 1.2 M lithium chloride, 500 mM Tris-HCl, and 8 M urea (pH 8.0). Adducts were visualized with a Storm 860 Imager (Molecular Dynamics, Sunnyvale, CA). HPLC/Electrospray Ionization Mass Spectrometry. Online sample preparation and cleanup were performed using two automated TPMV switching valves (Rheodyne) and two HPLC pumps. Switching valve A allowed HPLC pump A (Dionex GP40, Dionex, Sunnyvale, CA) to either load and wash or backflush the trap column. Switching valve B diverted the trap column effluent either to a waste reservoir or to the analytical column. HPLC pump A was used for sample cleanup, analysis, and column cleaning. HPLC pump B, a syringe-type pump (ISCO 260D, ISCO, Lincoln, NE), was used to equilibrate the analytical column with 35% acetonitrile and to maintain flow through the analytical column and mass spectrometer during sample loading and preparation periods. Samples were loaded for 3 min onto the trap column (C18 Poros 10-R2; 2.1 mm × 30 mm; PerSeptive Biosystems, Framingham, MA) using water at a flow rate of 1 mL/min. The column was then washed for 3 min with 30% methanol at 1 mL/min, with the effluent directed to the waste reservoir. Both valves were subsequently switched, and the concentrated sample was loaded onto the analytical column (5 µm Nucleosil C18, 2 mm × 250 mm, Keystone Scientific, Bellefonte, PA; or 5 µm Columbus C18, 2 mm × 150 mm, Phenomenex, Torrance, CA) and into the mass spectrometer using 35% acetonitrile at a flow rate of 200 µL/min. After an 11 min analysis period, both the trap and analytical columns were eluted with a 5 min gradient to 90% acetonitrile, followed by a 5 min gradient to 35% acetonitrile. Both switches were then set to the initial position to allow reequilibration of both columns using pump B. Analyses were conducted with a VG Platform II singlequadrupole mass spectrometer (Micromass, Altrincham, England) equipped with an electrospray ionization interface. The ion source temperature was 150 °C, and positive ions were acquired in a selected ion monitoring mode (dwell time of 0.3 s, span of 0.02 Da, and interchannel delay time of 0.03 s) for the M + H+ and BH2+ ions of both dG-C8-4-ABP and dG-C8-4-ABPd9. The sampling cone-skimmer voltage was switched between 20 and 70 V to produce in-source collision-induced dissociation in concert with acquisition of the respective selected ion. The

Beland et al. M + H+ ions for dG-C8-4-ABP (m/z 435.2) and dG-C8-4-ABPd9 (m/z 444.2) were acquired at 20 V; the BH2+ ions (m/z 319.2 and 328.2, respectively) were acquired at 70 V. Dissociation-Enhanced Lanthanide Fluoroimmunoassay. The competitive dissociation-enhanced lanthanide fluoroimmunoassay (DELFIA) method has been described previously by Schoket et al. (20). The DELFIA for dG-C8-4-ABP employed an antiserum elicited against N-(guanosin-8-yl)-2-acetylaminofluorene (21, 22), microtiter plates coated with a human serum albumin (HSA) conjugate of N-(guanosin-8-yl)-2-acetylaminofluorene, and a standard curve prepared with G-C8-4ABP. The N-(guanosin-8-yl)-2-acetylaminofluorene was conjugated to HSA by periodate oxidation (21), and 220 pg of the conjugate was coated onto each microtiter plate well. Nonspecific binding was blocked with 1% fetal calf serum (FCS) in phosphatebuffered saline [140 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4‚ 7H2O, 1.4 mM KH2PO4, and 3.1 mM NaN3 (pH 7.3)] containing 0.05% Tween-20 (PBS/Tween, NIH Media Unit, Bethesda, MD) for 1 h at 37 °C. Rabbit anti-N-(guanosin-8-yl)-2-acetylaminofluorene (20, 21) was diluted 1:80000 with 2% FCS in PBS/ Tween, and the solution was mixed with either the standard competitor (G-C8-4-ABP) in the presence of digested nonmodified DNA or an equal volume of a digested DNA sample. The mixtures were incubated for 90 min at 37 °C in triplicate experimental wells and one control well coated only with HSA. The plates were then incubated for 90 min at 37 °C with biotinylated anti-rabbit IgG (Vector Laboratories, Burlingame, CA) that had been reconstituted in water (1.5 mg/mL) and diluted 1:2500 in PBS/Tween with 1% FCS. Europium-labeled streptavidin (Wallac, Gaithersburg, MD) was added, and the plates were incubated for 60 min at room temperature. Wallac enhancement solution was added, and the plates were further incubated for 60 min at room temperature. The fluorescence was then measured in a DELFIA 1234 research fluorometer (Wallac) at 614 nm. Fifty percent inhibition of the standard curves was observed at 42 ( 16 fmol of G-C8-4-ABP/well (mean ( SD; n ) 14). Since modified DNA is recognized inefficiently by this assay, the DNA samples were digested to either 3′-nucleotides with micrococcal nuclease and spleen phosphodiesterase (5) or nucleosides with DNase I, alkaline phosphatase, and snake venom phosphodiesterase (19) before the analyses. Aliquots of the enzyme digests were then assayed by DELFIA.

Results In Vitro Reaction of Calf Thymus DNA with N-Hydroxy-4-aminobiphenyl. [2,2′-3H]-N-Hydroxy-4aminobiphenyl was reacted with calf thymus DNA at molar ratios of 0, 1.8 × 10-5, 1.8 × 10-4, 1.8 × 10-3, and 1.8 × 10-2 N-hydroxy4-aminobiphenyl:DNA nucleotides. After purification of the DNA by solvent extractions and precipitation, the extent of adduct formation was determined by UV spectroscopy and scintillation counting. The binding levels, as determined by 3H incorporation, were 0, 62, 320, 6500, and 11 000 adducts/108 nucleotides, respectively (Table 1). Aliquots of the samples were enzymatically hydrolyzed to nucleosides and analyzed by HPLC. The two samples with the highest extent of modification (i.e., 6500 and 11 000 adducts/108 nucleotides; 200 µg aliquots) were analyzed directly. Since considerably larger amounts of DNA were required to analyze the two samples modified at lower levels (i.e., 62 and 320 adducts/108 nucleotides; 20 and 1 mg, respectively), the adducts were partitioned into n-butanol before analysis. The HPLC profiles from the two highly modified samples (6500 and 11 000 adducts/108 nucleotides) contained a major peak of radioactivity (40-50% of the radioactivity eluting through the column) that coeluted with the added marker, dGC8-4-ABP (Figure 1a). Additional radioactivity eluted at

4-Aminobiphenyl-DNA Adduct Standard

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 71

Table 1. Comparison of the Adduct Levels (Expressed as dG-C8-4-ABP/108 Nucleotides) in DNA Modified in Vitro with 4-Aminobiphenyl Using Different Detection Methodologies reaction ratioa 0.0 1.8 × 10-5 1.8 × 10-4 1.8 × 10-3 1.8 × 10-2

3H

incorporationb 0.0g

62 ( 0.8 320 6500 11000

32P-postlabelingc

0.42 0.84 (1.3%)i 3.0 (0.9%) 300 (4.6%) 480 (4.4%)

HPLC/ESI/MSd 0.0 19 ( 1.7 (31%) 56 ( 1.8 (18%) 4300 ( 210 (66%) nd

DELFIAe (MN/SPD) ndh

82 ( 26 (130%) nd nd nd

DELFIAf (DNase I/SVPD/AP) nd 63 ( 20 (100%) nd nd nd

a Molar ratio between N-hydroxy-4-aminobiphenyl and DNA nucleotides in the reaction. b The extent of binding as determined by scintillation counting after treating the DNA with DNase I. c The amount of dG-C8-4-ABP as determined by the extent of incorporation of 32P, assuming a specific activity of 3950 Ci/mmol for [γ-32P]ATP. d The amount of dG-C8-4-ABP as determined by HPLC/electrospray ionization mass spectrometry (HPLC/ESI/MS) by quantifying against a dG-C8-4-ABP-d9 internal standard. e The amount of dG-C8-4ABP as determined by DELFIA. The DNA was hydrolyzed to 3′-nucleotides with micrococcal nuclease (MN) and spleen phosphodiesterase (SPD). G-C8-4-ABP was used as the standard competitor. f The amount of dG-C8-4-ABP as determined by DELFIA. The DNA was hydrolyzed to nucleosides with DNase I, snake venom phosphodiesterase (SVPD), and alkaline phosphatase (AP). G-C8-4-ABP was used as the standard competitor. g The data are reported as a single measurement or the mean ( SD of at least three measurements. h Not determined. i The percentages are relative to the values determined by direct measurement of 3H incorporation in the same samples.

(Figure 1b). The sample modified at 320 adducts/108 nucleotides was also chromatographed by HPLC after extraction with n-butanol. Approximately 60% of the radioactivity partitioned into n-butanol, and a peak of radioactivity coeluting with the dG-C8-4-ABP adduct marker was observed (not shown). 32P-Postlabeling

Analyses of DNA Reacted in Vitro with N-Hydroxy-4-aminobiphenyl. 32P-Postlabeling analyses of all the calf thymus DNA samples reacted with [2,2′-3H]-N-hydroxy-4-aminobiphenyl showed a major DNA adduct (Figure 2b,c, spot 1) that did not occur in the solvent control sample (Figure 2a). The same adduct was observed when the 3′-nucleotide of dG-C84-ABP was labeled (not shown). The major adduct was accompanied by a second adduct with approximately 15% of the intensity of the former (Figure 2b,c, spot 2). If it is assumed that the [γ-32P]ATP had a specific activity of 3950 Ci/mmol (4), the adduct levels as determined by 32P incorporation were 0.9-4.6% of those indicated by the 3H measurements (Table 1).

Figure 1. HPLC analyses of the enzymatic hydrolysate from calf thymus DNA reacted with [2,2-3H]-N-hydroxy-4-aminobiphenyl at pH 5. The elution conditions are outlined in Materials and Methods. The adduct dG-C8-4-ABP, which was added to serve as a UV marker, eluted at 22 min. One minute fractions were collected to assess the amount of radioactivity. (a) Profile obtained with 200 µg of DNA from a reaction conducted at a molar ratio of 1.8 × 10-2:1 N-hydroxy-4-aminobiphenyl:DNA nucleotides. The enzymatic hydrolysate was analyzed directly. (b) Profile obtained with 20 mg of DNA from the reaction conducted at a 1.8 × 10-5:1 molar ratio. The adducts were extracted with n-butanol before HPLC analysis.

the beginning (ca. 9%) and end (ca. 10%) of the chromatogram. The former was probably due to incomplete digestion of the DNA. The identity of the latter is not known, but the long retention time suggests that it was associated with nonbonded lipophilic materials. Approximately 65% of the radioactivity associated with the sample containing 62 adducts/108 nucleotides partitioned into n-butanol. When the n-butanol fraction was analyzed by HPLC, two peaks of nearly equal intensity were observed, each accounting for 20% of the radioactivity eluting through the column. The second of the two peaks coeluted with the adduct marker, dG-C8-4-ABP

HPLC/Electrospray Ionization Mass Spectrometric Analyses of DNA Reacted in Vitro with N-Hydroxy-4-aminobiphenyl. The full scan mass spectrum of dG-C8-4-ABP, acquired with a sampling cone-skimmer voltage of 20 V, consisted of two predominant ions, M + H+ (m/z 435) and BH2+ (m/z 319; Figure 3a). Increasing the cone-skimmer voltage to 70 V decreased the intensity of the M + H+ ion and increased the intensity of the BH2+ ion (Figure 3b). Further increases in cone-skimmer voltages resulted in fragmentation of the BH2+ ion (Figure 3c). To maximize the signals, cone-skimmer voltages of 20 and 70 V were chosen for selected ion monitoring of the M + H+ and BH2+ ions, respectively. The same voltages were used for selected ion monitoring of the M + H+ (m/z 444) and the BH2+ (m/z 328) ions from the internal standard, dG-C8-4-ABP-d9. Calibration curves for the concentration ratio versus dG-C8-4-ABP:dG-C8-4-ABP-d9 response ratio were generated for both the M + H+ and BH2+ ions. Over a 0.17.6 range of concentration ratios, the response ratios with both ions were highly linear (r2 g 0.99), with slopes of 1.13 and 0.95 for the M + H+ and BH2+ ions, respectively. Structural confirmation of the dG-C8-4-ABP was obtained by monitoring for coelution of both the M + H+ and BH2+ ions; thus, it was possible to quantify dG-C84-ABP even if an interference appeared in one of the ion traces. The detection limit (signal/noise ) 3) was approximately 23 fmol.

72 Chem. Res. Toxicol., Vol. 12, No. 1, 1999

Beland et al.

Figure 2. 32P-Postlabeling analyses of calf thymus DNA reacted with [2,2-3H]-N-hydroxy-4-aminobiphenyl at pH 5. The elution conditions are outlined in Materials and Methods. (a) Profile obtained from the control reaction. (b) Profile obtained from the reaction conducted at a molar ratio of 1.8 × 10-5:1 N-hydroxy-4-aminobiphenyl:DNA nucleotides. (c) Profile obtained from the reaction conducted at a 1.8 × 10-2:1 molar ratio. The adducts were visualized with a phosphorimager using attenuation factors of 96, 96, and 5400, respectively, for panels a-c. Each analysis was conducted with 10 µg of DNA. Region 1 was used to quantify dG-C8-4-ABP after subtracting an equivalent background area. The region marked X in panel b was attributed to an impurity.

Figure 3. In-source collision-induced dissociation mass spectrometric analysis of dG-C8-4-ABP. Full scan mass spectra of 300 ng of dG-C8-4-ABP were obtained using electrospray ionization at cone-skimmer voltages of (a) 20, (b) 70, and (c) 100 V.

dG-C8-4-ABP-d9 was added to 50-100 µg aliquots of 4-aminobiphenyl-modified DNA at approximately the anticipated level of dG-C8-4-ABP. The samples were enzymatically hydrolyzed to nucleosides, and aliquots were analyzed directly by HPLC/electrospray ionization mass spectrometry. The average recovery of the dG-C84-ABP-d9 standard through the entire procedure ranged from 50 to 80%. The use of d0/d9 response ratios compensated for losses due to differences in recovery and changes in the response of the instrument. Figure 4 presents typical chromatograms. Figure 4a shows the results from a 1:1 mixture of dG-C8-4-ABP and dG-C8-4-ABP-d9, which eluted at 8.9-9.0 min. Selected

ions for the M + H+ ions of dG-C8-4-ABP-d9 (m/z 444) and dG-C8-4-ABP (m/z 435) are presented in panels a1 and a2 of Figure 4, respectively; the selected ions for the BH2+ ions of dG-C8-4-ABP-d9 (m/z 328) and dG-C8-4-ABP (m/z 319) are given in panels a3 and a4 of Figure 4. Figure 4b shows the same ions for the control DNA sample spiked with 117 pg of dG-C8-4-ABP-d9. The M + H+ and BH2+ ions of dG-C8-4-ABP-d9 are readily evident (panels b1 and b3 of Figure 4), while the analogous ions from dG-C8-4-ABP are clearly absent (panels b2 and b4 of Figure 4). Figure 4c illustrates the results from the reaction conducted at a molar ratio of 1.8 × 10-3:1 N-hydroxy-4-aminobiphenyl:DNA nucleotides spiked with

4-Aminobiphenyl-DNA Adduct Standard

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 73

Figure 4. Selected ion chromatograms of M + H+ and BH2+ of calf thymus DNA reacted with [2,2-3H]-N-hydroxy-4-aminobiphenyl at pH 5. (a) Profiles obtained from the dG-C8-4-ABP and dG-C8-4-ABP-d9 standards. (b) Profiles obtained from the control reaction. (c) Profiles obtained from the reaction conducted at a molar ratio of 1.8 × 10-3:1 N-hydroxy-4-aminobiphenyl:DNA nucleotides. Selected ions for the M + H+ ions of dG-C8-4-ABP-d9 (m/z 444) and dG-C8-4-ABP (m/z 435) were obtained at cone-skimmer voltages of 20 V. Selected ions for the BH2+ ions of dG-C8-4-ABP-d9 (m/z 328) and dG-C8-4-ABP (m/z 319) were obtained at cone-skimmer voltages of 70 V. dG-C8-4-ABP-d9 (117 and 253 pg for panels b and c, respectively) was added at the beginning of the enzymatic hydrolysis. dG-C8-4-ABP and dG-C8-4-ABP-d9 eluted at 8.9-9.0 min.

253 pg of dG-C8-4-ABP-d9. The M + H+ and BH2+ ions of dG-C8-4-ABP-d9 (panels c1 and c3 of Figure 4) and dGC8-4-ABP (panels c2 and c4 of Figure 4) are clearly present. The results from the HPLC/mass spectral analyses are summarized in Table 1. The adduct levels were 18-66% of those determined by the 3H measurements, with a relative standard deviation of 5-9% (n ) 3-7). DELFIA of DNA Reacted in Vitro with N-Hydroxy-4-aminobiphenyl. The DNA sample modified in vitro to the lowest extent with 4-aminobiphenyl (i.e., 62 adducts/108 nucleotides as determined by 3H incorporation) was assayed by a newly developed competitive DELFIA. Since the assay is specific for dG-C8-4-ABP as a monomer, the DNA samples were digested before analysis. Two different hydrolysis procedures were used. In one, the DNA was digested to 3′-nucleotides using conditions similar to those used in 32P-postlabeling; in the other, the DNA was digested to nucleosides using the same conditions used for the HPLC analyses. The levels of dG-C8-4-ABP and its 3′-phosphate as determined by the DELFIA were 63 ( 20 and 82 ( 26 adducts/108 nucleotides (mean ( SD; n ) 6 and 6), respectively, values comparable to those indicated by the 3H measurements (Table 1). Analyses of DNA Modified in Vivo with 4-Aminobiphenyl. B6C3F1 mice were given a single intraperi-

toneal injection of [2,2′-3H]-4-aminobiphenyl at doses of 0.1 and 1.0 mg per mouse. Hepatic DNA was isolated, and the extent of binding was assessed by measuring the extent of 3H incorporation using liquid scintillation counting, which indicated a 6.3-fold difference in the extent of adduct formation between the two doses (Table 2). Due to the limited amount of radioactivity associated with the hepatic DNA (250-1500 dpm/mg of DNA), these samples were not characterized by HPLC. Aliquots of the liver DNA were analyzed by 32P-postlabeling using conditions identical to those described for the in vitro modified DNA. As with the in vitro samples, a major adduct was observed in the hepatic DNA samples (Figure 5). The adduct levels, as determined by 32P incorporation, were 3.4-5.4% of those found with the 3H measurements (Table 2). When analyzed by HPLC/electrospray ionization mass spectrometry, the hepatic DNA from the mice treated with 4-aminobiphenyl had a major peak, dG-C84-ABP, which eluted at 9 min. This peak was not observed in mice treated with the solvent alone (Figure 6). The adduct levels, which were determined through comparison to the dG-C8-4-ABP-d9 internal standard, were 61-70% of the values (n ) 3) indicated by the 3H measurements (Table 2), presumably due to incomplete enzymatic hydrolysis of the DNA, losses of the adduct during chromatography, or both. The recovery of dG-C84-ABP-d9 through the entire procedure was 70-90%.

74 Chem. Res. Toxicol., Vol. 12, No. 1, 1999

Beland et al.

Table 2. Comparison of the Adduct Levels (Expressed as dG-C8-4-ABP/108 Nucleotides) in DNA Modified in Vivo with 4-Aminobiphenyl Using Different Detection Methodologies in vivo samplea analysis method

0.0 mg

0.1 mg

1.0 mg

incorporationb 32P-postlabelingd HPLC/ESI/MSf DELFIA (MN/SPD)g DELFIA (DNase I/SVPD/AP)i

0.0c 0.07 0.0 ndh nd

70 3.8 (5.4%)e 49 ( 5.0 (70%) 180 ( 56 (260%) 180 ( 66 (260%)

440 15 (3.4%) 270 ( 15 (61%) 1700 ( 660 (390%) 2400 ( 690 (550%)

3H

a B6C3F1 mice were treated intraperitoneally with 0, 0.1, or 1.0 mg of [2,2′-3H]-4-aminobiphenyl. After 24 h, the mice were killed by exposure to CO2, livers were pooled by group, and DNA was prepared. b The extent of binding as determined by scintillation counting after treating the DNA with DNase I. c The data are reported as a single measurement or the mean ( SD of at least three measurements. d The amount of dG-C8-4-ABP as determined by the extent of incorporation of 32P, assuming a specific activity of 3950 Ci/mmol for [γ-32P]ATP. e The percentages are relative to the values determined by direct measurement of 3H incorporation in the same samples.f The amount of dG-C8-4-ABP as determined by HPLC/electrospray ionization mass spectrometry (HPLC/ESI/MS) by quantifying against a dG-C8-4-ABP-d9 internal standard. g The amount of dG-C8-4-ABP as determined by DELFIA. The DNA was hydrolyzed to 3′-nucleotides with micrococcal nuclease (MN) and spleen phosphodiesterase (SPD). G-C8-4-ABP was used as the standard competitor. h Not determined. i The amount of dG-C8-4-ABP as determined by DELFIA. The DNA was hydrolyzed to nucleosides with DNase I, snake venom phosphodiesterase (SVPD), and alkaline phosphatase (AP). G-C8-4-ABP was used as the standard competitor.

Figure 5. 32P-Postlabeling analyses of hepatic DNA from mice treated intraperitoneally with (a) 0.0, (b) 0.1, or (c) 1.0 mg of [2,23H]-4-aminobiphenyl. The adducts were visualized with a phosphorimager using an attenuation factor of 500. Each analysis was conducted with 10 µg of DNA.

DELFIA analyses were conducted after hydrolyzing the hepatic DNA to either 3′-nucleotides or nucleosides. Both hydrolysis conditions gave similar results, with the adduct levels being considerably higher (260-550%) than those determined by direct measurement of the 3H associated with the DNA (Table 2).

Discussion Initial attempts to prepare a DNA standard modified with 4-aminobiphenyl were based upon reactions with N-acetoxy-N-(trifluoroacetyl)-4-aminobiphenyl (12, 13). Satisfactory results were obtained when preparing highly adducted DNA or when modifying deoxyguanosine or guanosine; however, our goal was to prepare a DNA standard modified at levels approaching those found in vivo, and under these conditions, the solvolysis of N-acetoxy-N-(trifluoroacetyl)-4-aminobiphenyl appeared to be favored over its reaction with DNA. For this reason, modifications were conducted using N-hydroxy-4-aminobiphenyl at a slightly acidic pH, a condition that facilitates DNA adduct formation from N-hydroxyarylamines (23). The major adduct resulting from this reaction is dGC8-4-ABP (24-26). Reactions were conducted over a 1000-fold range in N-hydroxy-4-aminobiphenyl concentrations, which re-

sulted in a 180-fold range in the extent of binding as measured by the amount of 3H associated with the DNA (Table 1). HPLC analysis of the more highly modified samples indicated primarily one major adduct that coeluted with the added marker, dG-C8-4-ABP (Figure 1a). With the less modified samples, an additional peak was observed that eluted before the marker (Figure 1b). This early peak had a retention time nearly identical to that of G-C8-4-ABP, which suggested that it could have arisen from reaction with residual RNA in the calf thymus DNA. Through comparison to an authentic standard, mass spectral analyses demonstrated that this was not the case (data not shown). The material could be a second adduct; however, for reasons presented below, we believe it was a degradation product of N-hydroxy4-aminobiphenyl that was not removed efficiently. Two 32P-postlabeling procedures, based on nuclease P1 sensitivity (4) and n-butanol extraction (5), are typically used to assay aromatic DNA adducts. Since the nuclease P1 sensitivity method tends to dephosphorylate adducts substituted at C8 of deoxyguanosine (27, 28), the analyses were conducted using the n-butanol extraction procedure. Each of the in vitro modified DNA samples gave one major adduct (Figure 2) accompanied by smaller amounts of other adducts. This adduct pattern was also found

4-Aminobiphenyl-DNA Adduct Standard

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 75

Figure 6. Selected ion chromatograms of M + H+ and BH2+ of hepatic DNA from mice treated intraperitoneally with [2,2-3H]-4aminobiphenyl. (a) Profiles obtained from the dG-C8-4-ABP and dG-C8-4-ABP-d9 standards. (b) Profiles obtained from hepatic DNA from mice treated with solvent. (c) Profiles obtained from hepatic DNA from mice treated with 1.0 mg of [2,2-3H]-4-aminobiphenyl. Selected ions for the M + H+ ions of dG-C8-4-ABP-d9 (m/z 444) and dG-C8-4-ABP (m/z 435) were obtained at cone-skimmer voltages of 20 V. Selected ions for the BH2+ ions of dG-C8-4-ABP-d9 (m/z 328) and dG-C8-4-ABP (m/z 319) were obtained at cone-skimmer voltages of 70 V. dG-C8-4-ABP-d9 (117 pg) was added at the beginning of the enzymatic hydrolysis. The dG-C8-4-ABP and dG-C84-ABP-d9 eluted at 8.8-9.0 min.

when these samples were analyzed by other laboratories (29), and has been observed previously (26, 30) with other samples. With the samples modified at low levels (i.e., from reactions conducted with molar ratios of 1.8 × 10-5 and 1.8 × 10-4 N-hydroxy-4-aminobiphenyl:DNA nucleotides), the yield of dG-C8-4-ABP based upon the extent of 32P incorporation was only 0.9-1.3% of that anticipated from the 3H measurements (Table 1). This may be due to poor hydrolysis of the DNA, poor labeling of the adduct, loss of the adduct upon chromatography, or a combination of these factors. For example, in kinetic 32P-postlabeling studies conducted with 3′-nucleotides of deoxyguanosine and dG-C8-4-ABP, we have found that the apparent specificity constant (i.e., Kcat/Km) is approximately 4-fold lower for the adduct compared to that for deoxyguanosine 3′-phosphate.2 On the basis of the 3H measurements, the efficiency of 32P-postlabeling was approximately 4-fold better with the more highly modified samples (i.e., from reactions conducted with molar ratios of 1.8 × 10-3 and 1.8 × 10-2 N-hydroxy-4-aminobiphenyl:DNA nucleotides) compared to that with the samples modified at lower levels. This could be a function of the higher level of modification; for example, a decreased adduct labeling efficiency has been reported as adduct levels were 2 L. L. G. Mourato, F. A. Beland, and M. M. Marques, manuscript in preparation.

lowered (8). It is also consistent, however, with the interpretation that the additional “adduct” peak detected by HPLC (Figure 1b) is due to noncovalently bound material rather than it being an adduct. This interpretation is also compatible with the results obtained by HPLC/electrospray ionization mass spectrometry. At a reaction ratio of 1.8 × 10-3, the level of dG-C8-4-ABP was 66% of that indicated by the 3H measurements (Table 1), whereas at lower levels of modification, the values detected by mass spectrometry were only 18-31% of the anticipated level. These results could be due to a 2-4fold lower efficiency of ion detection for samples modified at low levels as compared to those modified at higher levels. However, such a discrepancy in ion detection efficiency is inconsistent with the highly linear responses obtained by HPLC/electrospray ionization mass spectrometry when using the response ratio for the dG-C84-ABP and dG-C8-4-ABP-d9 standards over a similar range of concentrations. Therefore, an overestimation of the amount of dG-C8-4-ABP based upon the extent of 3H incorporation is likely to have occurred for the DNA samples modified at lower levels. Preliminary DELFIA results indicated that the antiserum recognized the 4-aminobiphenyl-DNA adduct in nonhydrolyzed DNA with an efficiency of only 10-15%. To increase the efficiency of adduct detection, aliquots

76 Chem. Res. Toxicol., Vol. 12, No. 1, 1999

from the DNA sample containing 62 adducts/108 nucleotides (based upon 3H incorporation) were hydrolyzed enzymatically to nucleosides or 3′-nucleotides. The conditions used to obtain nucleosides were identical to those used to hydrolyze the DNA for HPLC and HPLC/ electrospray ionization mass spectrometric analyses; the conditions used to obtain 3′-nucleotides were identical to those used in the 32P-postlabeling analyses. Control experiments with nonmodified calf thymus DNA indicated that the enzymatic digestions did not affect the 50% inhibition value or the slope of the G-C8-4-ABP standard curves. The DELFIA results with both hydrolysis conditions were similar to those obtained by direct measurement of the 3H associated with the DNA (Table 1). Because the HPLC and the HPLC/electrospray ionization mass spectrometry data indicated that the 3H measurements overestimated the amount of dG-C8-4-ABP, and the 32P-postlabeling results indicated that no significant amounts of other adducts were present, the level of dGC8-4-ABP indicated by DELFIA is difficult to explain. One possibility is that the antiserum has a higher affinity for dG-C8-4-ABP than for the G-C8-4-ABP standard. Should this be the case, the DELFIA would overestimate the extent of dG-C8-4-ABP modification in DNA. This interpretation is supported by additional experiments in which a DNA sample highly modified with dG-C8-4-ABP (14 000 adducts/108 nucleotides as determined by 3H incorporation) was assayed by DELFIA, which indicated a binding level of 61 000 ( 19 000 adducts/108 nucleotides (mean ( SD; n ) 12). The goal of this study was to develop a DNA standard modified at a low adduct level (i.e., approximately 100 adducts/108 nucleotides) that could be used to quantify adduct formation from 4-aminobiphenyl in vivo. To test the utility of the standard, mice were treated with two different doses of [2,2′-3H]-4-aminobiphenyl, and hepatic DNA was isolated. On the basis of the level of 3H incorporation, the extent of binding (70 and 440 adducts/ 108 nucleotides for the low and high doses, respectively; Table 2) was comparable to the values obtained for the in vitro samples modified at low levels. When the in vivo samples were examined by 32P-postlabeling (Figure 5), the adduct profile was nearly identical to that observed with the in vitro modified samples (Figure 1). The level of 32P incorporation was similar to that observed with the more highly modified in vitro samples (i.e., 3-5% of the value anticipated from the 3H measurements), which further supports the conclusion that the in vitro samples modified at low levels contained a substantial amount of nonbonded 3H. The presence of dG-C8-4-ABP in the in vivo samples was confirmed by HPLC/electrospray ionization mass spectrometry, which indicated that adduct levels were 60-70% of those found by the 3H measurements (Table 2), fully consistent with the percentage obtained for the more highly modified in vitro sample analyzed by HPLC/electrospray ionization mass spectrometry. These results contrast with those obtained by DELFIA, which overestimated the amount of dG-C84-ABP in the in vivo samples by 260-550%. This overestimation of adduct levels by DELFIA is probably due to the antiserum recognizing the G-C8-4-ABP standard with a lower affinity than that with which it detects dG-C8-4-ABP. As noted above, when the concentration of dG-C8-4ABP in the in vivo samples was quantified using the amount of 32P incorporated, the adduct values were

Beland et al. Table 3. Quantification of Adduct Levels (Expressed as dG-C8-4-ABP Adducts/108 Nucleotides) in DNA Modified in Vivo with 4-Aminobiphenyl Based upon the in Vitro DNA Standarda in vivo sampleb analysis method 32P-postlabeling

0.1 mg

1.0 mg

(120%)c,d

86 340 (77%) DELFIA (MN/SPD) 42 ( 13 (60%) 400 ( 150 (91%) DELFIA (DNase I/SVPD/AP) 54 ( 20 (77%) 720 ( 210 (160%)

a The level of dG-C8-4-ABP in the in vivo samples was estimated using the in vitro modified standard and assuming an adduct level of 19 adducts/108 nucleotides, as determined by HPLC/electrospray ionization mass spectrometry. b B6C3F1 mice were treated intraperitoneally with 0, 0.1, or 1.0 mg of [2,2′-3H]-4-aminobiphenyl. After 24 h, the mice were killed by exposure to CO2, livers were pooled by group, and DNA was prepared. c The data are reported as a single measurement or the mean ( SD of at least three measurements. d The percentages are relative to the values determined by direct measurement of 3H incorporation in the same samples.

severely underestimated compared to the levels indicated by the 3H measurements (Table 2). To rectify this deficiency, we assessed the usefulness of the in vitro modified DNA to serve as a quantification standard. A satisfactory estimation of in vivo adduct levels was achieved by 32P-postlabeling analyses through comparison to the standard, assuming that the level of dG-C84-ABP in the standard was 19 adducts/108 nucleotides, as determined by HPLC/electrospray ionization mass spectrometry (Table 3). Similarly, when the DELFIA assays were quantified using the value of 19 adducts/ 108 nucleotides for the in vitro standard, the results were quite satisfactory (Table 3).

Conclusions We have prepared DNA containing dG-C8-4-ABP that can be used as a standard for 32P-postlabeling and other DNA adduct detection methodologies. The presence of dG-C8-4-ABP was confirmed by HPLC and HPLC/electrospray ionization mass spectrometry. 32P-Postlabeling analyses, based upon the extent of incorporation of 32P, consistently underestimated the levels of dG-C8-4-ABP, while the DELFIA, using a G-C8-4-ABP quantitation standard, overestimated the adduct levels. The adduct levels determined by HPLC/electrospray ionization mass spectrometry best reflected those obtained from the extent of 3H incorporation. By using the in vitro DNA adduct standard, there was a better correspondence among all three adduct detection methodologies (32Ppostlabeling, HPLC/electrospray ionization mass spectrometry, and DELFIA) when assaying dG-C8-4-ABP in in vivo samples. This standard is available from the authors upon request.

Acknowledgment. We thank Fred F. Kadlubar for providing the purified calf thymus DNA and Cindy Hartwick for helping prepare the manuscript. B. Schoket was supported, in part, by the U.S.-Hungarian Science and Technology Research Fund (J.F.363).

References (1) Randerath, K., Reddy, M. V., and Gupta, R. C. (1981) 32P-Labeling test for DNA damage. Proc. Natl. Acad. Sci. U.S.A. 78, 61266129. (2) Gupta, R. C., Reddy, M. V., and Randerath, K. (1982) 32Ppostlabeling analysis of non-radioactive aromatic carcinogen-DNA adducts. Carcinogenesis 3, 1081-1092.

4-Aminobiphenyl-DNA Adduct Standard (3) Beach, A. C., and Gupta, R. C. (1992) Human biomonitoring and the 32P-postlabeling assay. Carcinogenesis 13, 1053-1074. (4) Reddy, M. V., and Randerath, K. (1986) Nuclease P1-mediated enhancement of sensitivity of 32P-postlabeling test for structurally diverse DNA adducts. Carcinogenesis 7, 1543-1551. (5) Gupta, R. C. (1985) Enhanced sensitivity of 32P-postlabeling analysis of aromatic carcinogen:DNA adducts. Cancer Res. 45, 5656-5662. (6) Reddy, M. V., Gupta, R. C., Randerath, E., and Randerath, K. (1984) 32P-Postlabeling test for covalent DNA binding of chemicals in vivo: application to a variety of aromatic carcinogens and methylating agents. Carcinogenesis 5, 231-243. (7) Randerath, E., Agrawal, H. P., Weaver, J. A., Bordelon, C. B., and Randerath, K. (1985) 32P-Postlabeling analysis of DNA adducts persisting for up to 42 weeks in the skin, epidermis and dermis of mice treated topically with 7,12-dimethylbenz[a]anthracene. Carcinogenesis 6, 1117-1126. (8) Shields, P. G., Harris, C. C., Petruzzeli, S., Bowman, E. D., and Weston, A. (1993) Standardization of the 32P-postlabeling assay for polycyclic aromatic hydrocarbon-DNA adducts. Mutagenesis 8, 121-126. (9) Phillips, D. H., and Castegnaro, M. (1993) Results of an interlaboratory trial of 32P-postlabelling. In Postlabelling Methods for Detection of DNA Adducts (Phillips, D. H., Castegnaro, M., and Bartsch, H., Eds.) IARC Scientific Publication 124, pp 35-49, International Agency for Research on Cancer, Lyon, France. (10) Gupta, R. C. (1993) 32P-Postlabeling analysis of bulky aromatic adducts. In Postlabelling Methods for Detection of DNA Adducts (Phillips, D. H., Castegnaro, M., and Bartsch, H., Eds.) IARC Scientific Publication 124, pp 11-23, International Agency for Research on Cancer, Lyon, France. (11) Westra, J. G. (1981) A rapid and simple synthesis of reactive metabolites of carcinogenic aromatic amines in high yield. Carcinogenesis 2, 355-357. (12) Lee, M.-S., and King, C. M. (1981) New syntheses of N-(guanosin8-yl)-4-aminobiphenyl and its 5′-monophosphate. Chem.-Biol. Interact. 34, 239-248. (13) Marques, M. M., and Beland, F. A. (1990) Synthesis, characterization, and solution properties of ras sequences modified by arylamine carcinogens at the first base of codon 61. Chem. Res. Toxicol. 3, 559-565. (14) Crivello, J. V. (1981) Nitrations and oxidations with inorganic nitrate salts in trifluoroacetic anhydride. J. Org. Chem. 46, 30563060. (15) Marques, M. M., Mourato, L. L. G., Santos, M. A., and Beland, F. A. (1996) Synthesis, characterization, and conformational analysis of DNA adducts from methylated anilines present in tobacco smoke. Chem. Res. Toxicol. 9, 99-108. (16) Shapiro, R., Underwood, G. R., Zawadzka, H., Broyde, S., and Hingerty, B. E. (1986) Conformation of d(CpG) modified by the carcinogen 4-aminobiphenyl: a combined experimental and theoretical analysis. Biochemistry 25, 2198-2205. (17) Basler, J., Hastie, N. D., Pietras, D., Matsui, S.-I., Sandberg, A. A., and Berezney, R. (1981) Hybridization of nuclear matrix attached deoxyribonucleic acid fragments. Biochemistry 20, 69216929. (18) Beland, F. A., Fullerton, N. F., and Heflich, R. H. (1984) Rapid isolation, hydrolysis and chromatography of formaldehyde-modi-

Chem. Res. Toxicol., Vol. 12, No. 1, 1999 77 fied DNA. J. Chromatogr. 308, 121-131. (19) Heflich, R. H., Morris, S. M., Beranek, D. T., McGarrity, L. J., Chen, J. J., and Beland, F. A. (1986) Relationships between the DNA adducts and the mutations and sister-chromatid exchanges produced in Chinese hamster ovary cells by N-hydroxy-2-aminofluorene, N-hydroxy-N′-acetylbenzidine and 1-nitrosopyrene. Mutagenesis 1, 201-206. (20) Schoket, B., Doty, W. A., Vincze, I., Strickland, P. T., Ferri, G. M., Assennato, G., and Poirier, M. C. (1993) Increased sensitivity for determination of polycyclic aromatic hydrocarbon-DNA adducts in human DNA samples by dissociation-enhanced lanthanide fluoroimmunoassay (DELFIA). Cancer Epidemiol., Biomarkers Prev. 2, 349-353. (21) Poirier, M. C., Yuspa, S. H., Weinstein, I. B., and Blobstein, S. (1977) Detection of carcinogen-DNA adducts by radioimmunoassay. Nature 270, 186-188. (22) Poirier, M. C., and Connor, R. J. (1982) Radioimmunoassay for 2-acetylaminofluorene-DNA adducts. Methods Enzymol. 84, 607618. (23) Kadlubar, F. F., and Beland, F. A. (1985) Chemical properties of ultimate carcinogenic metabolites of arylamines and arylamides. In Polycyclic Hydrocarbons and Carcinogenesis (Harvey, R. G., Ed.) ACS Symposium Series 283, pp 341-370, American Chemical Society, Washington, DC. (24) Kadlubar, F. F., Beland, F. A., Beranek, D. T., Dooley, K. L., Heflich, R. H., and Evans, F. E. (1982) Arylamine-DNA adduct formation in relation to urinary bladder carcinogenesis and Salmonella typhimurium mutagenesis. In Environmental Mutagens and Carcinogens (Sugimura, T., Kondo, S., and Takebe, H., Eds.) pp 385-396, Alan R. Liss, New York. (25) Beland, F. A., Beranek, D. T., Dooley, K. L., Heflich, R. H., and Kadlubar, F. F. (1983) Arylamine-DNA adducts in vitro and in vivo: their role in bacterial mutagenesis and urinary bladder carcinogenesis. Environ. Health Perspect. 49, 125-134. (26) Talaska, G., Dooley, K. L., and Kadlubar, F. F. (1990) Detection and characterization of carcinogen-DNA adducts in exfoliated urothelial cells from 4-aminobiphenyl-treated dogs by 32P-postlabelling and subsequent thin-layer and high-pressure liquid chromatography. Carcinogenesis 11, 639-646. (27) Gupta, R. C., and Earley, K. (1988) 32P-adduct assay: comparative recoveries of structurally diverse DNA adducts in the various enhancement procedures. Carcinogenesis 9, 1687-1693. (28) Gallagher, J. E., Jackson, M. A., George, M. H., Lewtas, J., and Robertson, I. G. C. (1989) Differences in detection of DNA adducts in the 32P-postlabelling assay after either 1-butanol extraction or nuclease P1 treatment. Cancer Lett. 45, 7-12. (29) Castegnaro, M., and Phillips, D. H. (1997) Interlaboratory standardisation and validation of DNA adduct postlabelling methods for human studies. Final Report on E.U. Project No. EV5V-CT940448. International Agency for Research on Cancer, Lyon, France. (30) Poirier, M. C., Fullerton, N. F., Smith, B. A., and Beland, F. A. (1995) DNA adduct formation and tumorigenesis in mice during the chronic administration of 4-aminobiphenyl at multiple dose levels. Carcinogenesis 16, 2917-2921.

TX980172Y