Polyclonal Antibodies to Adenine N1-Oxide: Characterization and Use

Measurement of oxidatively generated base damage in cellular DNA. Jean Cadet , Thierry Douki , Jean-Luc Ravanat. Mutation Research/Fundamental and ...
0 downloads 0 Views 122KB Size
Chem. Res. Toxicol. 1998, 11, 1169-1175

1169

Polyclonal Antibodies to Adenine N1-Oxide: Characterization and Use for the Measurement of DNA Damage Nathalie Signorini, Didier Molko, and Jean Cadet* De´ partement de Recherche Fondamentale sur la Matie` re Condense´ e, Service de Chimie Inorganique et Biologique, Laboratoire des Le´ sions des Acides Nucle´ iques, CEA/Grenoble, 17 avenue des Martyrs, F-38054 Grenoble Cedex 9, France Received March 5, 1998

Adenine N1-oxide is a DNA lesion whose formation involves the specific oxidation of the adenine base by hydrogen peroxide under nonradical conditions. The damage may be measured using a HPLC/32P-postlabeling method, which however cannot be used for routine analysis. We propose herein as an alternative an immunological assay which allows a rapid evaluation of the level of adenine N1-oxide in DNA exposed to oxidative stress. Two polyclonal antibodies were raised using two different strategies for the coupling of the hapten to the protein. The first approach is based on the universal method of Erlanger and Beiser, whereas the preparation of the second antigen involves the conjugation of a morpholino derivative of adenosine N1oxide to the carrier protein. The affinity and the specificity of those antibodies were determined by competitive enzyme-linked immunosorbent assay. The antibody obtained by the traditional method shows some cross-reactivity with normal nucleotides, whereas for the other antiserum, the selectivity was found to be higher. Therefore, this polyclonal antibody was used to quantify the level of adenine N1-oxide in calf thymus DNA oxidized either by m-chloroperbenzoic acid or by hydrogen peroxide. The detection limit of the assay is four residues of adenine N1-oxide per 106 normal bases. The level of adenine N1-oxide in nonmodified DNA was lower than the detection limit of the assay, whereas in mCPB- and H2O2-modified DNA, it could be up to 14 and 0.7 adenine N1-oxide residues per 104 normal bases, respectively.

Introduction Various systems (ionizing radiation, ultraviolet and solar lights, and reactive oxygen species) can induce oxidative damage to genomic DNA, including base modifications, DNA strand breaks, abasic sites, and DNAprotein cross-links (1-3). It is now well-established that oxidative DNA lesions are implicated in several deleterious biological effects, including aging, mutagenesis, carcinogenesis, and cell lethality (4). Therefore, the quantitative measurement of modified purine and pyrimidine bases is essential in efforts to better delineate the role of DNA damage in oxidative stress. Such a determination constitutes a challenging analytical problem; it has to be specific and sensitive enough to be used for the measurement of damage to cellular DNA. Several analytical approaches, including 32P-postlabeling assays (5-7), immunological methods (8, 9), HPLC in combination with electrochemical detection, gas chromatography coupled to mass spectrometry (10, 11), and capillary electrophoresis (12, 13), have been developed for the detection and the measurement of DNA base modifications (for comprehensive reviews, see refs 14 and 15). Immunological detection has long been recognized potentially as a powerful tool for the analysis of modified DNA by genotoxics. Since antibodies were raised by * To whom correspondence should be addressed. Phone: (33) 4 76 88 49 87. Fax: (33) 4 76 88 50 90. E-mail: [email protected].

Levine et al. (16) against far-UV-induced photoproducts of DNA, several immunoassays have been designed for measuring oxidative and photoinduced DNA damage. However, to our knowledge, no antibody against adenine N1-oxide, one form of oxidative base damage to DNA, was available. Adenine N1-oxide is a specific product of the reaction of hydrogen peroxide with the adenine moiety of DNA under nonradical conditions (17). It was shown that exposure of fibroblast cells to a broad source of near-UV light (340-420 nm) induced the formation of adenine N1oxide (18). Mouret et al. (6) developed a 32P-postlabeling assay to monitor the formation of the latter lesion within cellular DNA exposed to hydrogen peroxide. The sensitivity of the HPLC/32P-postlabeling method is close to one modification per 106 normal bases. However, the assay is not convenient for routine analysis, since it is a tedious method. It should be mentioned that a capillary zone electrophoresis assay, which allowed the measurement of the level of adenine N1-oxide within DNA, recently became available (19). However, the method suffers from a lack of sensitivity since the limit of detection is three adenine N1-oxide residues per 105 normal nucleotides. Therefore, this explains why attempts were made to design a more sensitive method, which should be easy to use in to measuring the level of adenine N1-oxide in cellular DNA. We report herein the development and the characterization of polyclonal antibodies that recognize adenine N1-oxide in DNA. Using these antibodies in the sensitive

10.1021/tx980044+ CCC: $15.00 © 1998 American Chemical Society Published on Web 10/02/1998

1170 Chem. Res. Toxicol., Vol. 11, No. 10, 1998

enzyme-linked immunoassay (ELISA1), it was possible to readily detect four lesions per 106 normal bases.

Materials and Methods Chemicals. Adenosine, m-chloroperbenzoic acid (mCPB), sodium periodate, sodium borohydride, and N′-[(3-dimethylamino)propyl]-N-ethylcarbodiimide hydrochloride (EDC) were purchased from Aldrich Chemical Co. (Milwaukee, WI). Adenosine N1-oxide (AdoNox), glycine, bovine serum albumin (BSA), ovalbumin (OVA), methylated bovine serum albumin (mBSA), 3,3′,5,5′-tetramethylbenzidine (TMB), H2O2, and Tween 20 were supplied by Sigma Chemical Co. (St. Louis, MO). Ethylene glycol and formic acid were obtained from Merck (Darmstadt, Germany). Polystyrene 96-well plates were purchased from Dynatech (Polylabo, Strasbourg, France). Phosphate-buffered saline (PBS) was made up of 140 mM NaCl, 2.7 mM KCl, 1.5 mM KH2PO4, and 10 mM Na2HPO4 (pH 7.4). Water was purified with a Milli-Q system (Milford, MA). Cytidine N3-oxide was synthesized as described by Delia et al. (20). High-Performance Liquid Chromatography. For system A, a Merck Li-Chrospher 100RP-18 end-capped 5 µm silica gel (125 mm × 4 mm) column was used and the eluent was 25 mM triethylammonium acetate (TEAA, pH 7) at 1 mL/min. For system B, a Li-Chroprep RP-18 15-25 µm silica gel (150 mm × 40 mm) column from Merck was used with an axial compression system from Jobin-Yvon (Lonjumeau, France). The elution was achieved with 100% water. The flow rate was controlled by a Moduloprep pump (Jobin-Yvon), and compounds were detected with a SM 25 UV spectrophotometer (Delsy-Nermag, Argenteuil, France). For system C, a Nucleosil 7C-18 (7 µm, 250 mm × 21 mm) Macherey-Nagel column (Du¨ren, Germany) was used and the eluent was 100% H2O. For system D, “flash chromatography”, a C-18 reversed phase column (Econosil prep 90, Alltech) was used. The elution was carried out with 25 mM TEAA (pH 7). Capillary Electrophoresis. Capillary zone electrophoresis (CZE) analyses were performed with a Beckman P/ACE 5500 System, coupled to an UV detector (214 nm) and controlled by System Gold software. An untreated fused-silica capillary (57 cm total length x 75 µm i.d.) was used. The temperature of the analytical system was maintained at 18 °C. The samples (∼5 nL) were injected under pressure for 1 s. The running solution was 50 mM borate buffer adjusted to pH 10 ( 0.1 with a 1 M NaOH solution. The instrument was set at a fixed voltage of 25 kV, leading to a constant current of 66 µA. The quantitative analysis of the four normal nucleotides was made using their corrected peak areas. Spectroscopic Analysis. FAB mass spectra were recorded in the positive mode on a ZAB2-SEQ spectrometer (Fisons-V. G., Manchester, U.K.) equipped with a LSIMS source. The molecules were dissolved in a glycerol matrix and desorbed by cesium ion bombardment (35 keV). The 400.13 MHz 1H NMR and the 100.61 13C NMR spectra were recorded in the Fourier transform mode on a Bruker 400 instrument (Bruker, Wissembourg, France). The 1H and 13C chemical shifts are expressed in parts per million with respect to 3-(trimethylsilyl)propionic acid (TSP) used as the internal reference in 99.99% deuterium oxide. According to nomenclature rules, the anomeric proton is labeled H1′ in the morpholinonucleoside series. 1 Abbreviations: EDC, N′-[(3-dimethylamino)propyl]-N-ethylcarbodiimide hydrochloride; BSA, bovine serum albumin; OVA, ovalbumin; mBSA, methylated bovine serum albumin; TMB, 3,3′,5,5′-tetramethylbenzidine; PBS, phosphate-buffered saline; CMAHM, 4-(carboxymethyl)-2-(adenin-9-yl)-6-(hydroxymethyl)morpholine; CMHMANox, 4-(carboxymethyl)-2-(adenine N1-oxide-9-yl)-6-(hydroxymethyl)morpholine; Ado, adenosine; AdoNox, adenosine N1-oxide; dAdo, 2′-deoxyadenosine; dAdoNox, 2′-deoxyadenosine N1-oxide; dAMP5′, 2′-deoxyadenosine 5′-monophosphate; dANoxMP5′, 2′-deoxyadenosine N1-oxide 5′-monophosphate; mCPB, m-chloroperbenzoic acid; TEAA, triethylammonium acetate; ELISA, enzyme-linked immunosorbant assay.

Signorini et al. Synthesis of 4-(Carboxymethyl)-2-(adenin-9-yl)-6-(hydroxymethyl)morpholine (CMHMA). CMHMA was synthesized from adenosine following the procedure which is described below for CMHMANox. The purification was achieved by HPLC using system B (yield of 18%): 200.13 MHz 1H NMR (D2O, TSP) δ 2.42 (pseudotriplet, 1H, H-3′), 2.93 (pseudotriplet, 1H, H-2′), 3.08 (pseudotriplet, 1H, H-3′′), 3.27 (singlet, 2H, CH2), 3.36 (pseudotriplet, 1H, H-2′′), 3.79 (multiplet, 2H, H-5′ and H-5′′), 4.19 (multiplet, 1H, H-4′), 5.98 (quadruplet, 1H, H-1′), 8.21 (singlet, 1H, H-2), 8.35 (singlet, 1H, H-8); 100.61 MHz 13C NMR (D2O, TSP) δ 52.3 (C-5′), 54.7 (CH2), 61.0 (C-2′), 62.1 (C-3′), 76.5 (C-4′), 79.3 (C-1′), 118.0 (C-5), 139.8 (C-8), 147.9 (C-4), 152.5 (C-2), 155.2 (C-6), 177.1 (COOH); FAB-MS (positive mode, relative intensity) m/z 331 [91, (M + Na)+], 309 [42, (M + H)+], 174 [30, (morpholine moiety)+], 136 [100, (B + 2H)+]. Synthesis of 4-(Carboxymethyl)-2-(adenine N1-oxide-9yl)-6-(hydroxymethyl)morpholine (CMHMANox). Adenosine N1-oxide (500 mg) was placed in water (40 mL). Although the nucleoside was not completely dissolved, sodium periodate (NaIO4) was added (416 mg, 1.1 equiv), and then the mixture became homogeneous. After the mixture was stirred for 20 min at room temperature, the excess NaIO4 was decomposed by addition of ethylene glycol (22 µL, 0.22 equiv), and the solution was kept for 5 min at room temperature. Then, 5 equiv (664 mg) of glycine was added, and the pH was adjusted to 9.0-9.5 with a 5% solution of potassium carbonate. After 45 min, sodium borohydride (168 mg, 2.5 equiv) was added, and the reaction mixture was set aside for 18 h. The reaction was stopped by decreasing the pH to 7.0-8.0 using 1 M formic acid. The resulting residue was then purified by flash chromatography using system D. The homogeneity of the collected fractions of interest was checked using system A. Three fractions that contained the desired product with different degrees of purity (80, 55, and 30%) were obtained. Then, CMHMANox was purified by preparative HPLC using system C (yield of 35%): 400.13 MHz 1H NMR (D2O, TSP) δ 2.51 (pseudotriplet, 1H, H-3′), 3.04 (pseudotriplet, 1H, H-2′), 3.10 (pseudotriplet, 1H, H-3′′), 3.33 (singlet, 2H, CH2), 3.41 (pseudotriplet, 1H, H-2′′), 3.84 (multiplet, 2H, H-5′ and H-5′′), 4.23 (multiplet, 1H, H-4′), 6.13 (quadruplet, 1H, H-1′), 8.58 (singlet, 1H, H-8), 8.72 (singlet, 1H, H-2); 100.61 MHz 13C NMR (D2O, TSP) δ 52.2 (C-5′), 54.6 (CH2), 61.0 (C-3′), 62.0 (C-2′), 76.6 (C4′), 79.6 (C-1′), 118.7 (C-5), 142.9 (C-8), 143.2 (C-4), 144.3 (C-2), 148.6 (C-6), 177.1 (COOH); FAB-MS (positive mode, relative intensity) m/z 347 [65, (M + Na)+], 325 [100, (M + H)+], 174 [35, (morpholine moiety)+], 152 [52, (B + 2H)+]. Synthesis of 2′-Deoxyadenosine N1-Oxide 5′-Monophosphate. 2′-Deoxyadenosine 5′-monophosphate (dAMP5′, 300 mg) was dissolved in 120 mL of phosphate buffer (Tritisol, Merck, pH 7). One hundred milliliters of a 0.01 M ethanolic solution of m-chloroperbenzoı¨c acid (mCPB) (1.3 equiv) was added, and the reaction mixture was kept for 18 h at room temperature. Then, the solution was concentrated to dryness, and the resulting white residue was dissolved in 50 mL of water. Subsequently, mCPB was extracted with 100 mL of diethyl oxide (three times). The aqueous solution was concentrated, and its content was purified by flash chromatography using system D (yield of 51%). After purification of dANoxMP5′, the triethylammonium counterions of the phosphate groups were exchanged for sodium ions using an ion-exchange resin (Dowex-50W-X8, Na+ form, 100-200 mesh). Preparation of Protein-Nucleoside Conjugates Using the Method of Erlanger and Beiser. Adenosine N1-oxideBSA and adenosine N1-oxide-OVA (OVA-AdoNox) conjugates were prepared following the protocol described by Erlanger and Beiser (24). The amount of hapten bound to the proteins was determined using the UV absorption features of the protein, the hapten, and the conjugate. The relative contributions of the nucleoside and BSA were estimated by measuring the absorbance at 280 and 260 nm, respectively. It was found that ten and five residues of adenosine N1-oxide were linked to BSA and ovalbumin, respectively.

Immunodetection of Adenine N1-Oxide Residues Preparation of Protein-Nucleoside Conjugates According to the Morpholinonucleoside Method. CMHMANox (6 mg dissolved in 0.7 mL of water) was added to a solution of mBSA (16.9 mg, 0.013 equiv) in 2 mL of H2O containing EDC (7.1 mg, 2 equiv) over a period of 3 h. The reaction mixture was stirred at room temperature for 18 h. Then, the mixture was dialyzed against H2O and finally lyophilized. The hapten: carrier ratio was determined as described above using the UV absorption method. There were about 20 CMHMANox residues linked to mBSA. Production of the Antibodies. Two rabbits were immunized with the adenosine N1-oxide-BSA conjugate, whereas two other rabbits were immunized with CMHMANox coupled to mBSA. The immunogen (1 mg in 1 mL of water) was emulsified with Freund’s complete adjuvant (1 mL) and subcutaneously injected at multiple sites on the shaved backs of the animals. The same procedure was applied again 4 weeks later. Then, 8 and 12 weeks after the beginning of the protocol, each rabbit received one intramuscular injection of the same amount of antigen in complete Freund’s adjuvant. Two weeks after the last immunization, the rabbits were bled from the carotid artery, yielding the maximun amount of serum (protocol conducted by the Elevage Scientifique des Dombes, Chaˆtillon-sur-Chalaronne, France). Characterization of the Antibodies. Optimal conditions for ELISA tests were determined using a checkerboard procedure in which coating antigen levels of 0.5 ng/well to 10 µg/well and antiserum dilutions of 1:10 to 1:106 were tested. With regard to the serum prepared according to the method of Erlanger and Beiser (Erlanger-Beiser antiserum), an adequate absorbance was found using 10 ng of OVA-AdoNox conjugate/ well and an antiserum dilution of 1:104. On the other hand, for the morpholinonucleoside type serum (morpholinonucleoside antiserum), the same optimal conjugate concentration was found but the antiserum dilution was 1:103. The specificity of each antiserum was then determined using a competitive ELISA protocol. Typically, 96-well polystyrene plates were coated with 10 ng of the OVA-AdoNox conjugate diluted in 40 µL of PBS and dried at 37 °C overnight. Plates could be stored for up to 2 months at 4 °C under dry and lightprotected conditions without any detectable loss of performance. The plates were washed three times with PBS/Tween 20 (0.05%) using a platewasher (LP35, Sanofi Diagnostics Pasteur, Marnes-La-Coquette, France) and then dried by beating onto absorbent paper towels. Nonspecific binding sites were blocked by incubation in 3% (w/v) dry skim milk in PBS for 1 h at room temperature, and the plates were washed three times with PBS/Tween 20 (0.05%). Then, standard solutions of inhibitors were diluted in PBS so that the concentrations varied between 10-1 and 107 fmol/well. Subsequently, 25 µL of each of the latter solutions was pipetted onto the plate in rows (four wells). Usually, two rows on the ELISA plate were used for positive controls. In those wells, the inhibitor was substituted by 25 µL of PBS. A reference row containing PBS (50 µL/well) was also used. Polyclonal rabbit antiserum (25 µL) was added to each well in the plate with the exception of those in the reference row. Then, the plates were incubated for 90 min in darkness at room temperature. After the washing of the plates (four times) with PBS/Tween 20, a 50 µL sample of the second antibody, goat anti-mouse IgG peroxidase conjugate, was added to each well at a 1:25000 dilution, and the plates were further incubated at room temperature in the absence of light for 90 min. The incubation was followed by five washings in PBS/Tween 20 and a final washing with distilled water. Finally, 50 µL of the substrate solution [1 mg of TMB in 0.1 mL of DMSO and 2 µL of 30% H2O2 in 10 mL of citrate buffer (pH 5.0)] for the peroxidase-conjugated second antibody was added in each well. After incubation for 7 min at room temperature, under light-protected and airless conditions, the reaction was stopped with 50 µL of 1 M HCl, and absorbances were read at 450 nm with a Labsystems Multiskan plate reader (Helsinki, Finland).

Chem. Res. Toxicol., Vol. 11, No. 10, 1998 1171 DNA Modification Using mCPB. Calf thymus DNA (1.5 mg) was dissolved in water (2.5 mL), and the salts were removed on a Pharmacia NAP-25 column. In a subsequent step, six aliquots were prepared from the latter solution and then heated at 90 °C for 5 min prior to being chilled on ice. At this stage, 200 µL of phosphate buffer (Tritisol, pH 7, Merck) was added to each tube. Then, DNA samples were modified with various amounts of a 0.1 M ethanolic solution of mCPB (0, 0.1, 0.2, 0.3, 0.5, and 0.6 equiv) for 30 min at room temperature. Extraction of mCPB was achieved with 1 mL of diethyl ether (three times), and the salts were removed on a Pharmacia NAP-10 column. The solutions were concentrated, and then the DNA samples were digested by nuclease P1 for 90 min at 37 °C. The exact concentration of 2′-deoxyribonucleoside 5′-monophosphates contained in the six tubes was estimated by capillary zone electrophoresis analyses using external calibration. DNA Modification using H2O2. Calf thymus DNA (150 µg) was dissolved in water (0.5 mL), and the salts were removed on a Pharmacia NAP-5 column. Then, four aliquots were prepared from the latter solution, and DNA was modified with various concentrations of H2O2 (0, 5 × 10-4, 10-3, and 10-2 M) for 30 min in darkness. The solutions were concentrated, and DNA was then digested by nuclease P1 for 90 min at 37 °C. The exact concentration of the 2′-deoxyribonucleoside 5′-monophosphates contained in the six tubes was determined by capillary zone electrophoresis analyses using external calibration.

Results and Discussion One of the main analytical approaches for monitoring the formation of oxidized bases within DNA is based on immunological detection. Such assays allow the measurement of oxidative and photoinduced DNA damage within either isolated DNA (21, 22) or individual cells (16, 23). However, the assays require antibodies which are highly specific and devoid of any significant crossreactivity with the overwhelming normal bases. In this study, a sensitive ELISA approach was used both to characterize the antibody and to determine the level of adenine N1-oxide in oxidized DNA. Preparation of the Antigens. Two approaches were used to bind the morpholinonucleoside of adenine N1oxide to the carrier protein. On one hand, the universal method of Erlanger and Beiser was used, and the corresponding serum was called the Erlanger-Beiser antiserum (24). On the other hand, we synthesized and purified by HPLC the morpholinonucleoside of adenine N1-oxide. Then, the latter compound was linked to a carrier protein with EDC; the corresponding serum was called morpholinonucleoside antiserum (25). In this case, the purification of the hapten occurred just before its coupling to the carrier protein. Initially, the morpholinonucleoside of adenine (CMAHM) was synthesized from adenosine (Ado); then, after purification, CMAHM was oxidized with mCPB to obtain CMHMANox. The FAB mass spectrometry analysis provided evidence for the incorporation of one oxygen atom in the molecule. Furthermore, the product exhibited a strong UV absorption at 229 nm which is suggestive of the presence of an N-oxide group. However, the 1H NMR results were puzzling, since the chemical shifts of the protons of the morpholine moiety showed large disparities compared with those of CMAHM. This is strongly suggestive of the occurrence of mCPB-induced oxidation within the morpholine moiety. To overcome this difficulty, a slightly different strategy was used. Following the same approach used by Girault

1172 Chem. Res. Toxicol., Vol. 11, No. 10, 1998

Signorini et al.

Figure 1. Preparation of the conjugate adenosine N1-oxide linked to mBSA, via the morpholinonucleoside method. Table 1. Cross-Reactivity of Two Different Polyclonal Antibodies concentration for 50% inhibition (pmol/well) inhibitor

Figure 2. Inhibition curve for dAdo and dAdoNox in competitive ELISA.

et al. (26) for raising antibodies against 8-oxo-7,8-dihydro2′-deoxyguanosine, CMHMANox was prepared directly from adenosine N1-oxide (Figure 1). This was achieved by condensing the 2′,3′-dialdehyde derivative of adenosine N1-oxide with glycine. For this purpose, reaction conditions used by Rayford et al. (27) were optimized; adenosine N1-oxide was oxidized with sodium periodate, and the resulting dialdehyde was condensed with glycine. The double Schiff base was stabilized by reduction using sodium borohydride, instead of sodium cyanoborohydride as defined initially by Rayford et al. After purification, the modified morpholinonucleoside was coupled to the protein. About 20 residues of nucleosides were found to be linked to mBSA. It should be noted that the level of binding was higher than that obtained with the Erlanger-Beiser conjugate (five residues of nucleosides per mBSA). Characterization of the Antibodies. Preliminary experiments were performed to optimize the ELISA parameters for the detection of adenine N1-oxide. For each antiserum, a checkerboard procedure was used to establish conditions for optimal color development. It was found that this required a 10 ng coating of antigen per well. In addition, the working dilution was 1:104 and 1:103 for the Erlanger-Beiser antiserum and the morpholinonucleoside antiserum, respectively. Thus, the antisera were tested in a competitive ELISA to check whether they recognized the adenine N1-oxide lesion. Both antisera were inhibited by dAdoNox; meanwhile, dAdo did not inhibit them as shown in Figure 2. These observations strongly suggest that the antisera contain antibodies against the adenine N1-oxide lesion.

CMHMANox dAdoNox dANoxMP5′ AdoNox adenine N1-oxide cytosine N3-oxide dAdo dAMP5′ adenosine adenine FAPY adenine CMHMA N1-methyl-2′-deoxyadenosine N-methylmorpholine morpholinonucleoside of thymine 8-oxo-dGuo 2′-deoxyinosine a

Erlanger-Beiser morpholinoserum nucleoside serum 0.6 1 2.1 3.5 9.8 >97000a >19000 >100000 >52000a >6700a >41000 13000 680000 >2700000 >150000a

0.8 0.8 1.3 1.2 2.6 >97000 >19000 >100000a >52000a >6700a >41000 32000 110000 >2700000a >150000a

>11000a >50000a

>11000a >50000a

At least 70% of control (noninhibition).

Antibody specificities were determined in a competitive ELISA assay using compounds structurally related to dAdoNox. The results are summarized in Table 1. With regard to the Erlanger-Beiser antiserum, the most potent inhibitor was CMHMANox, which was 2-, 3.5-, 6-, and 16-fold more efficient than dAdoNox, dANoxMP5′, AdoNox, and adenine N1-oxide, respectively. In the case of the morpholinonucleoside antiserum, CMHMANox and dAdoNox gave similar results. It should be added that they were 1.5-, 1.6-, and 3-fold more efficient than AdoNox, dANoxMP5′, and adenine N1-oxide, respectively. Thus, for the Erlanger-Beiser antiserum, the sugar moiety seemed to be more important than for the morpholinonucleoside antiserum. In contrast, the antiserum from the morpholinonucleoside appeared to recognize antigenic determinants located on the altered base rather than those on the sugar. It could be noticed that CMAHM inhibited both antibodies, about 10000-fold less for the ErlangerBeiser antiserum and 30000-fold less for the morpholinonucleoside antiserum than the N-oxide residues. This is not surprising since in both cases, the antigenic molecule was CMHMANox linked to mBSA. N1-Methyl2′-deoxyadenosine, a compound in which a methyl group was substituted for N-oxide, was about 700000- and 100000-fold less efficient than the N-oxide residues in inhibiting antigen-antibody binding with regard to Erlanger-Beiser and morpholinonucleoside antisera, respectively. This is in agreement with a previous hypothesis implying that the antigenic determinants were

Immunodetection of Adenine N1-Oxide Residues

Figure 3. Inhibition curves for a mixture of dANoxMP5′ and DNA digested by nuclease P1 (two dANoxMP5′ residues per 104 normal nucleotides): (a) Erlanger-Beiser antiserum and (b) morpholinonucleoside antiserum.

located on the base rather than on the sugar for the morpholinonucleoside antiserum. In contrast to the Erlanger-Beiser antiserum, both the base and sugar were important. In each competitive assay, the concentrations of adenine, corresponding nucleosides, and nucleotides were the most elevated that we were able to obtain. So we can conclude that the selectivity was higher than one dANoxMP5′ per 4.7 × 104 normal nucleotides for the Erlanger-Beiser antiserum and higher than one dANoxMP5′ per 7.7 × 104 normal nucleotides for the morpholinonucleoside antiserum. We used a mixture of hydrolyzed DNA and 2′-deoxyadenosine N1-oxide 5′-monophosphate (dANoxMP5′) as the inhibitor to investigate more deeply the cross-reactivity of the antisera. First, dANoxMP5′ was added to DNA so as to obtain two residues of dANoxMP5′ per 104 normal nucleotides. Then, the solution was diluted 50% with PBS. The dilutions were used to plot an inhibition curve ranging from 9.1 to 0.07 pmol of dANoxMP5′ per well (Figure 3). It should be noted that the standard deviation bars on all the figures represent three measurements of one sample. Since the response was linear for both antisera, an additional experiment was carried out using a solution of 0.7 dANoxMP5′ per 104 normal nucleotides. The curve reported in Figure 4 was linear for the morpholinonucleoside antiserum. However, for the Erlanger-Beiser antiserum, the response was linear only until up to 4.55 pmol of dANoxMP5′ per well. Thus, it may be concluded that the antibodies recognized a chemical structure other than dANoxMP5′. A likely hypothesis is that the presence of an interfering substance in the DNA prevents the dANoxMP5′ from binding to the antibody in such a way that the antibody can now bind to the antigen on the plate. Such a cross-reactivity was expected for the ErlangerBeiser antiserum, since no purification of the hapten was done prior to the coupling to the protein. Furthermore, it may be mentioned that about 10% of the 2′-deoxyadenosine N1-oxide was converted to 2′-deoxyadenosine under the conditions of Erlanger and Beiser. Thus, both

Chem. Res. Toxicol., Vol. 11, No. 10, 1998 1173

Figure 4. Inhibition curves for a mixture of dANoxMP5′ and DNA digested by nuclease P1 (0.7 dANoxMP5′ residue per 104 normal nucleotides): (a) Erlanger-Beiser antiserum and (b) morpholinonucleoside antiserum.

Figure 5. Calibration curve obtained with competitive ELISA. Five micrograms of DNA digested by the nuclease P1, containing various amounts of dANoxMP5′ as the competitor (10 adenine N1oxide residues per 104 normal bases to four adenine N1-oxide residues per 106 normal bases).

adenine N1-oxide and adenine residues could be linked to protein under the conditions of Erlanger and Beiser. Therefore, purification of the hapten prior to its coupling to the carrier protein appears to be a requisite for improving the specificity of the antibodies. From these observations, it may be concluded that the Erlanger-Beiser antiserum could not be used in a sensitive immunoassay due to significant cross-reactivity with adenine. Therefore, only the morpholinonucleoside antiserum was applied to monitor the formation of adenine N1-oxide within DNA. Quantitative Analysis. To calibrate the measurement of adenine N1-oxide within DNA, known amounts of the dANoxMP5′ standard were mixed with nonoxidized DNA. Typically, to a fixed amount of immobilized antigen (adenosine N1-oxide-OVA) were added 5 µg of digested DNA (nuclease P1), various amounts of dANoxMP5′ as a competitor, and a fixed amount of antiserum. The antibody bound to the immobilized antigen was then determined and the amount compared to the amount bound in the absence of the competitor. The extent of the inhibition of antibody binding to the immobilized antigen was found to be proportional to the amount of competitor in the assay. With regard to the calibration curve, shown in Figure 5, an accurate response was achieved for hydrolyzed DNA

1174 Chem. Res. Toxicol., Vol. 11, No. 10, 1998

Signorini et al.

Table 2. Calf Thymus DNA Modified by Various Amounts of m-Chloroperbenzoic Acid for 30 min amount of mCPB used to modify calf thymus DNA (equiv) no. of dANoxMP5′ residues per 104 normal nucleotides

0

0.1

0.2

0.3

0.5

0.6