Aflatoxin B1 Dihydrodiol Antibody - PubMed Central Canada

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1984, p. 472-477 0099-2240/84/030472-06$02.00/0 Copyright © 1984, American Society for Microbiology

Vol. 47, No. 3

Aflatoxin B1 Dihydrodiol Antibody: Production and Specificity JAMES J. PESTKAt AND FUN S. CHU*

Food Research Institute and Department of Food Microbiology and Toxicology, University of Wisconsin, Madison, Wisconsin 53706 Received 26 October 1983/Accepted 13 December 1983

A specific antibody for 2,3-dihydro-2,3-dihydroxyaflatoxin B1 (AFB1-diol) was prepared, and its reactivity was characterized for the major aflatoxin (AF) B1 (AFB1) metabolites. Reductive alkylation was used to conjugate AFB,-diol to ethylenediamine-modified bovine serum albumin (EDA-BSA) and horseradish peroxidase for use as an immunogen and an enzyme-linked immunosorbent assay (ELISA) marker, respectively. High reactant ratios, 1:5 and 1:10, for AFB1-diol-EDA-BSA (wt/wt) resulted in precipitated conjugates which were poorly immunogenic. However, a soluble conjugate obtained by using a 1:25 ratio of AFB1-diol to EDA-BSA could be used for obtaining high-titer AFBI-diol rabbit antibody within 10 weeks. Competitive ELISAs revealed that the AFB,-diol antibody detected as little as 1 pmol of AFB1-diol per assay. Cross-reactivity of AFB1-diol antibody in the competitive ELISA with AF analogs was as follows: AFB1-diol, 100%; AFB1, 200%; AFM1, 130%; AFB2a, 100%; AFGI, 6%; AFG2, 4%; aflatoxicol, 20%; AFQ1, 2%; AFB1-modified DNA, 32%; and 2,3-dihydro-2-(N7-guanyl)-3-hydroxy AFB1, 0.6%. These data indicated that the cyclopentanone and methoxy moieties of the AF molecule were the primary epitopes for the AFB1-diol antibody. The AFB1-diol competitive ELISA was subject to substantial interference by human, rat, and mouse serum albumins but not by BSA, Tris, human immunoglobulin G, or lysozyme. By using a noncompetitive, indirect ELISA with an AFB1-modified DNA solid phase, a modification level of one AFB1 residue for 200,000 nucleotides could be determined.

rabbit reticulocyte systems and that relative rates of conversion of AFB1 to AFB1-diol in microsomal preparations from various species parallel the rank order of acute in vivo hepatotoxic effects in these animals. On this basis, it was suggested that the protein-binding ability of AFB1-diol is an important contributory factor in aflatoxicosis. However, at this time there is no conclusive evidence that AFBI-diol is similarly formed in vivo. The ability to detect AFBI-diol and related metabolites is therefore of fundamental importance to understanding the molecular mode of action of AFB1. Extracts and hydrolysates containing AFB, metabolites are typically analyzed by high-pressure chromatography or by use of radioisotopes. AFB1-diol presents a special problem because it binds to protein and therefore cannot be extracted. To overcome this, in vitro studies include Tris buffer, which contains primary amino groups that trap AFB1-diol immediately on formation. This allows easy extraction and quantitation of the AFB1-diol-Tris adduct by high-pressure liquid chromatography (22, 23). Recently we reported the efficacy of using specific antiserum prepared against 2,3-dihydro-2-hydroxy AFB1 (AFB2a) for the enzyme-linked immunosorbent detection of AFB1, AFB2a, AFB1-diol, AFBI-modified DNA, AFB1-N7-Gua, and AFB1-FaPyr (27). Production of the antiserum was based on the ability of AFB2a, like AFB,-diol, to react with primary amines at neutral and alkaline pH (1). A reductive alkylation method was thus used for conjugating the toxin to a bovine serum albumin (BSA) carrier protein for preparation of an immunogen. Resultant antibodies prepared against AFB2a-BSA had greatest specificity for the cyclopentanone and methoxy groups of the parent AF molecule (9, 28). AFB2a antibody has also been used in the indirect immunoperoxidase localization of bound AFB1 in rat liver (26). Since AFB1-diol is hydroxylated at the 2,3 position and AFB2a is hydroxylated only at the 2 position, we hypothesized that antiserum prepared against AFBI-diol conjugates

Aflatoxin (AF) B1 (AFB1) is a potent mycotoxin produced by certain strains of Aspergillus flavus and A. parasiticus that has been shown to occur naturally in human foods and animal feeds (5). Epidemiological studies in East Africa (25), the Phillipines (4), and Thailand (37) suggest a relationship between dietary AF intake and human liver cancer. AFB, and other AFs are also extremely hepatotoxic and have been implicated in a number of human and livestock toxicoses (31, 36). The carcinogenic and acute toxic effects of AFB, have been related to two highly reactive metabolites formed by hepatic microsomal enzymes. Metabolic activation of AFB1 by a cytochrome P450 to a reactive epoxide at position 2,3 of the terminal furan and subsequent covalent binding to nucleic acids have been suggested as key events in the carcinogenic process (21, 38). The primary DNA adduct formed in vitro and in vivo is 2,3-dihydro-2-(N7-guanyl)-3-hydroxy AFB1 (AFB,-N7-Gua) (8, 19, 20). Because of a localized positive charge, the imidazole ring of AFBI-N7-Gua can open under physiological conditions to yield a more stable, persistent adduct, the putative 2,3-dihydro-2-(N5-formyl-2',5',6'-triamino-4-oxo-N5-pyrimidyl)-3-hydroxy AFB, (AFB -FaPyr) (13, 19). Although covalent binding of the 2,3 epoxide to nucleophilic proteins is also likely, these adducts have not been characterized. A second major reactive metabolite, 2,3dihydro-2,3-dihydroxy AFB1 (AFBI-diol) is formed in vitro after spontaneous or enzymatic reaction of the 2,3 epoxide with water (18, 22, 23) and as a degradation product of

AFB1-N7-Gua-modified DNA (40). At physiological and

alkaline pH, AFB1-diol exists in a dialdehydic phenolate form and reacts with primary amino groups of proteins to form Schiff base adducts (22). Neal et al. (23) have demonstrated that AFB,-diol inhibits protein synthesis in isolated * Corresponding author. t Present address: Department of Food Science and Human Nutrition, Michigan State University, East Lansing, MI 48824-1224.

472

VOL. 47, 1984

might have even higher specificity and affinity for AFBI-diol and AFB1-modified DNA than one prepared against AFB2a. This study describes the preparation and characterization of AFB1-diol antibody and its application to detection of AFB diol and other key AF metabolites. (A preliminary report of this work was presented at the Annual Meeting of the American Society for Microbiology, 6 to 11 March 1983, New Orleans, La.) MATERIALS AND METHODS Materials. All inorganic chemicals and organic solvents were of reagent grade quality. Calf thymus DNA, BSA (fraction V, fatty acid-free, radioimmunoassay grade), horseradish peroxidase (type VI), lysozyme, human immunoglobulin G (IgG), human serum albumin (HSA; fraction V), mouse serum albumin (fraction V), rat serum albumin (fraction V), hydrogen peroxide, 2,2'-azino-di-3ethyl-benzthiazoline-6-sulfonate, and Tween 20 were purchased from Sigma Chemical Co., St. Louis, Mo. Male Fischer white rats (100 to 150 g) for liver S-9 and microsome preparations were obtained from Harlan Sprague-Dawley Laboratory, Indianapolis, Ind. New Zealand white rabbits were purchased from Klubertanz's Rabbit Farm, Edgerton, Wis., and determined to be Pasteurella negative before use. Preparation of AF derivatives. AFB1, AFB,, AFG1, AFG2, and AFM1 were purified from culture as previously described (6), and AFB1-diol was prepared by reacting AFBI with osmium tetroxide by the method of Swenson et al. (39). AFB2a and AFQ, were derived chemically by the procedures of Pohland et al. (32) and Buchi et al. (3), respectively. AFB1-N'-Gua DNA was prepared by the microsomal incubation procedure of Lin et al. (19) as modified by Pestka et al. (27). Preparation of immunogen and enzyme markers. For the protein carrier used in the immunogen, ethylenediamine (EDA) was conjugated to fatty acid-free BSA with watersoluble carbodiimide by the method of Chu et al. (7). AFB1diol was then conjugated to EDA-BSA by a modification of the reductive alkylation method previously used for AFB2a (9). Briefly, 0.05 to 1.0 ml of AFB1-diol in water (0.25 mg/ml) was added to 2 ml of EDA-BSA (2.5 mg/ml) in 0.05 M phosphate buffer (pH 7.2). The solution was gently swirled and allowed to stand for 30 min at 37°C. During this period the reaction mixture turned bright yellow. A 50-,lI volume of sodium borohydride solution (0.013 M) was added per ml of original reaction mixture and gently mixed, and the mixture was held for an additional 30 min at 25°C. A 25-,u volume of 0.1 N HCl per ml of reaction mixture was added to destroy residual NaBH4, and this mixture was dialyzed against 3 liters of 0.05 M phosphate buffer (pH 7.2) for 2 days (three changes) at 4°C. AFBI-diol concentrations and ratios of AFBI-diol-EDA-BSA conjugates were determined from the molar absorbance of AFB1 (e of 21,800 at 362 nm) (39). For the AFB,-diol-horseradish peroxidase ELISA marker, reductive alkylation was performed exactly as described for EDABSA, with a 1:50 (wt/wt) toxin-enzyme ratio. Immunization. Rabbits were immunized at multiple sites with 500 ,ug of immunogen (3 ml) mixed with saline-Freund complete adjuvant (1:2). Boosters of 500 ,ug of immunogen (1.5 ml) mixed with saline-Freund incomplete adjuvant (1:2) were injected intramuscularly at weeks 8 and 16. Rabbits were bled via marginal veins, and IgG was purified by the method of Herbert et al. (12). Titers were determined by ELISA as described by Pestka et al. (27). Competitive ELISA. The procedure for the microtiter plate competitive ELISA was identical to that described for

AFBI-DIOL ANTIBODY

473

AFB2a (28) except that AFB1-diol antibody and AFB,-diolperoxidase were used as described above for solid-phase immunosorbent and enzyme marker, respectively. Indirect ELISA. The ability of AFB1-diol and AFB2a antibodies to bind to AFB1-modified DNA was tested by a modification of the indirect ELISA of Hsu et al. (15). Portions (50 1d) of AFB1-adducted DNA or unmodified calf thymus DNA were air dried (40°C) onto the individual wells of polystyrene microtissue culture plates (no. 3040; BD Labware, Oxnard, Calif.), and the plates were washed three times with 0.1 M phosphate-buffered saline (PBS) (pH 7.5) containing 0.05% Tween 20 (PBS-Tween) on a Dynatech Miniwash (Dynatech Laboratories, Inc., Alexandria, Va.). Each well was filled with 0.25 ml of PBS containing 1% BSA (PBS-BSA), and the plates were incubated for 60 min at 37°C to minimize nonspecific binding. The plates were washed three more times and then incubated for 60 min at 37°C with 50 [LI of AFB1-diol antiserum diluted 1:1,000 in PBS-BSA. The wells were washed five times, and then the mixture was reacted with 50 p.l of goat anti-rabbit peroxidase conjugate (Miles Laboratories, Inc., Elkhart, Ind.), diluted 1:1,000 in PBS-BSA, for 30 min at 37°C. The plates were washed five more times, and the level of bound enzyme was determined as described by Pestka et al. (27). Absorbance at 414 nm was measured on a Dynatech Microelisa Minireader. RESULTS Immunogen preparation. Since AFB1-diol is a low-molecular-weight compound, it must be conjugated to a carrier protein to be rendered immunogenic. To facilitate this, EDA was first conjugated to free carboxyl groups of BSA to increase the number of available amino groups for subsequent mycotoxin conjugation (7). For conjugation of AFBrdiol to EDA-BSA, the two reactants were mixed in various ratios at pH 7.5 to form Schiff base adducts. Conjugates were stabilized by reduction with NaBH4. High ratios of mycotoxin to carrier protein, 1:5 (conjugate A) and 1:10 (conjugate B), resulted in precipitated conjugates during the first 30 min of reaction. (Absorbance at 362 nm could not be determined.). This was probably due to denaturation of the EDA-BSA after reaction with AFB1-diol. The lower reactant ratios, 1:25 (conjugate C), 1:50 (conjugate D), and 1:100 (conjugate E), yielded soluble conjugates with AFB1-diolEDA-BSA molar ratios of 7.4, 6.2, and 3.3, respectively. These final conjugate molar ratios are typical of those achieved previously for other mycotoxins (7). The final reaction volumes of conjugates A, B, C, D, and E were 6, 4, 2.8, 2.4, and 2.2 ml, respectively. Antibody production. To compare the relative immunogenicity of insoluble and soluble conjugates, we used conjugates A and C for antibody production in rabbits. Conjugate A, the insoluble immunogen prepared with a high toxin-tocarrier reactant ratio, failed to yield a significant antibody titer in either rabbit J-9 or J-10, reaching only 250 after 18 weeks (Table 1). However, conjugate C, the soluble immunogen with a final toxin-to-carrier molar ratio of 7.4, gave high titers, 3,240 and 9,720 for rabbits J-11 and J-12, respectively, after only 10 weeks. These results suggest that the toxin-carrier protein conjugates must be soluble for adequate antibody production. J-12 antiserum was used for all subsequent ELISA studies. Competitive ELISA and specificity. A competitive ELISA was used to test the suitability of the AFB1-diol antibody for detection of AFB,-diol and related analogs (Fig. 1). Linear response ranges were between 1 and 10 pmol per assay. Standard deviations among triplicate wells ranged between 3

PESTKA AND CHU

474

APPL. ENVIRON. MICROBIOL.

TABLE 1. ELISA titers for rabbits immunized with

EDA-BSA Wk

AFB,-diol-

ELISA titer in rabbit:

post-

immuni-

zationa

J_gb

J-10b

J-11

4 5 6 10 18

20 20 40 40

20 20

320 400 400 3,240 3,240

40 40 250

250

J-12c

640

1,600 800

9,720 3,240

Boosters given at weeks 8 and 16. b Immunized with insoluble conjugate A. c Immunized with soluble conjugate C. a

and 10%, typical of most ELISA procedures (24). To determine the specificity of the antiserum, we tested the ability of various AFB1 analogs to bind to AFB1-diol antibody in the competitive ELISA (Table 2). The results are expressed as the amount of AFBI analog required for 50% inhibition of AFB1-diol-peroxidase binding to the immunosorbent solid phase and are also compared with the specificity data for AFB2a antiserum (9, 27). It is apparent that AFB1 and AFM, were more effective than AFB1-diol in the competition. No differences were observed in competition by AFB2a or AFB1-diol in the AFB1-diol ELISA. The AFB1-diol antibody also had significant reactivity for AFB1-modified DNA and aflatoxicol and less reactivity for AFBI-N7-Gua, AFG1, AFG2, and AFQ1. The specificity of the antibody thus appeared to be directed to the cyclopentanone and methoxy moieties of the parent AFB1 molecule. Effect of protein on AFBI-diol competitive ELISA. Since the ability to detect AFB1-diol and related protein adducts would be useful in in vitro and in vivo AFB1 metabolism studies, the effect of preincubating AFB1-diol standards with Tris (0.5 M, pH 7.5) and various protein mixtures (5 mg/ml) at pH 7.5 for 1 h at 25°C on the standard AFB1-diol ELISA competition curve was tested. Under these conditions AFB1-diol should form Schiff base adducts with primary amino groups contained in the added compounds (22, 23). The amount of AFB1-diol required for 50% inhibition of 100-

c 75

m a

40

n0 E E

--

E

E 50 0

@

25-

O-1

0

1

2

log pmol/ assay

FIG. 1. Competitive ELISA for AFB1-diol. For the immunosorbent solid-phase, microtiter plates were coated with 0.05 ml of a 1:500 dilution of AFBI-diol antibody. AFBI-diol-peroxidase (8 ,ug/ ml) was incubated with increasing concentrations of AFB I-diol over the immunosorbent. Error bars indicate standard deviation.

TABLE 2. Specificity of AFB1-diol antiserum AFB1-diol antibody AFB2, antibody' 50% 50% Reactivity Reactivity AFB1 analog inhibition inhibition (% of (% of concn concn AFBI-diol AFB2a level) level) (pmol) (pmol) 3.2 100 AFB1-diol 1.0 80 AFB1 1.6 200 1.4 56 3.2 100 AFB2a 0.8 100 AFM1 2.5 130 76 1.1 AFG1 53 6 780 0.1 AFG2 80 4 780 0.1 Aflatoxicol 16 20 60 1.4 160 AFQ1 2 NDb ND AFB1-modified DNAC lod 32 1.7d 47 530d AFB1-N7-Gua 0.6 4.5d 18 a Data taken from previous work (9, 27). b ND, Not determined. C Unmodified DNA (up to 1,000 pmol per assay) did not demonstrably inhibit peroxidase conjugate binding in either assay. d Expressed as AFB1 concentration.

AFB,-diol-peroxidase binding to the immunosorbent phase

was again used as the standard for comparison. Preincubation with Tris, lysozyme, human IgG, and BSA had no effect on the standard AFBI-diol competition curve. However, the amount of AFB1-diol required for 50% inhibition increased from 3.2 pmol in the standard ELISA to 250 pmol when the compound was preincubated with HSA, mouse serum albumin, or rat serum albumin. When BSA was completely eliminated from the standard ELISA system by using carbonate buffer for coating the antibody or including 0.05% Tween 20 in all reaction mixtures to minimize nonspecific binding, similar results were found. The results indicate that the presence of albumin or components of these albumin fractions in serum and organ homogenates from humans, rats, and mice will interfere with the competitive immunosorption of AFB1-diol and AFBI-diol adducts in these fluids. Noncompetitive indirect ELISA for AFBt-modified DNA. The efficacy of using AFBI-diol antibody in a noncompetitive indirect ELISA for AFBl-modified DNA was tested. To determine optimal antiserum dilutions, AFB1-modified and unmodified calf thymus DNAs were air dried into wells of a polystyrene microtissue culture plate for preparation of the solid phase. Serial dilutions of AFBI-diol antiserum were incubated over the solid phase, and bound IgG was determined with a goat anti-rabbit peroxidase conjugate. The background binding of AFB1-diol antibody to calf thymus DNA was not significant in relation to AFB1-modified DNA (Fig. 2). When 50 ng of AFB1-modified DNA was used to coat the microplate wells, AFBl-diol antibody could be diluted 1:51,200 and still yield a signal-to-noise ratio (modified-unmodified) of 12. The effect of increasing the number of nucleotides relative to the number of AFB1 residues in the AFB1-diol noncompetitive ELISA for AFB1-modified DNA was determined (Fig. 3). The minimum detection limit for AFB1-modified DNA was arbitrarily defined as that level giving a signal-to-noise ratio of 2 or more. The results indicate that when 0.5 to 2.5 p.g of AFB1-modified DNA was added per well, a modification level of one AFB1 residue per 200,000 nucleotides could be determined. The addition of lower concentrations of AFB1-modified DNA, 25 and 5 ng per well, resulted in lower detection limits, one AFB1 residue in 1,600 nucleotides, and one AFB1 residue in 8,000 nucleotides, respectively.

AFB,-DIOL ANTIBODY

VOL. 47, 1984

0.9A Q8-

Q7-

0 c

05-

0-4.4

° 4 .0

0.3Q2

0.1-

0 2

3

4

I-

5

6

log serum dilution FIG. 2. Noncompetitive indirect ELISA of AFB1-modified DNA with AFB1-diol antibody. Wells were coated with 50 ,ul of AFB1modified or unmodified DNA. Modified DNA contained one AFB1 residue per 320 nucleotides. Serial dilutions of AFB1-diol antibody were incubated over the DNA solid phase, and bound antibody was determined with goat anti-rabbit peroxidase (1:1,000). Wells contained 50 (0), 10 (0), 5 (O), or 1 (E) ng of AFB1-modified DNA or 50 ng of unmodified calf thymus DNA (Ax).

DISCUSSION antibodies Specific against AFs and other carcinogens are extremely useful as immunological probes of the molecular mode of action of these compounds and as analytical tools for quantitating them in the environment. Several approaches for obtaining specific antibodies for AFB1 and its analogs have already been described by workers in our laboratory and by others (6, 9-11, 14, 16, 17, 29). We report here for the first time the conditions for preparing and characterizing AFB,-diol antisera, as well as the procedures for its use in AFIB1 metabolism studies. In regard to antibody preparation, useful antiserum could be obtained in only 10 weeks. A soluble AFB,-diol-EDA-BSA conjugate was essential for production of this high-titer antiserum. In general, insoluble conjugates are poor immunogens because insolubility renders themn inaccessible to the immunological system. In addition, the AFB1-diol molecule may also be masked because of the denaturation of the protein, making the toxin unavailable for immune stimulus. Although this might be generalized for other mycotoxin-carrier conjugates, it has been reported that T-2 toxin antibody can be more readily obtairied by immunization with insoluble toxin-protein conjugates rather than soluble ones (30). The AFB,-diol immunogen and the previously described AFB2a imniunogen are similar in that the cyclopentanone

475

ring and the methoxy group of the AF molecule acts as the primary epitope. The only difference between the two immunogens was the presence of a hydroxyl group at position 3 in AFB,-diol. Although it is likely that this hydroxyl group was located very close to the protein carrier and was not immunodominant, some differences in the relative specificities of the two antibodies were apparent. Table 2 compares the specifity data obtained for AFB1-diol antibody with those for AFB2a antibody (9, 27). The AFB1-diol antiserum appeared to have a broader range of specificity for AFB, analogs, especially aflatoxicol and AFM1, than did AFB2a antiserum. Whereas AFB, and AFM, were more effective than AFB,-diol and AFB2a in competing in the AFB1-diol ELISA, lower reactivities for AFB, and AFM, compared with those for AFB2a and AFB,-diol were found in the AFB2a ELISA. In general, the AFB1-diol antiserum required higher concentrations of AFB1-diol, AFB2a, AFB1-modified DNA, and AFB,-N7-Gua than did AFB2a antiserum to inhibit their respective peroxidase conjugates. This suggests that AFB2a antibody has higher affinity for these key metabolites than does AFB1-diol antibody. Although differences in specificity and affinity between these two antibodies may be related to the presence of an additional hydroxyl group on the AF molecule in the AFB1-diol-EDA-BSA immunogen, they may also be related to the presence of the EDA bridge. Further studies comparing the efficacy of EDA-BSA and BSA as carriers for AFB1-diol and AFB2a antiserum are therefore warranted. Also, since rabbit antisera vary between animals and with bleeding times, specific monoclonal antisera against AFB2a and AFB1-diol should be prepared for more thorough comparisons. It was noted that HSA, rat serum albumin, and mouse serum albumin, but not BSA, interfered with the AFB,-diol competitive ELISA. Noncovalent interactions between AFB1 and serum albumin have been previously demonstrated by electrophoresis (33, 41), column chromatography (34), spectrophotometric analysis (35), and equilibrium dialysis

1.2> 1.0we I. 0 c

h.

.0

IA0.2--~

2

3

4

5

6

7

log nucleotide:AFB1 ratio FIG. 3. Effect of nucleotide-AFB, substitution ratio on noncompetitive indirect ELISA for AFB1-modified DNA. Each well was coated with 50 ,ug of AFBI-modified DNA at various nucleotide-toAFB1 ratios, and binding of AFB1-diol antibody (1:1,000) was determined with goat anti-rabbit peroxidase. The ratios were varied by diluting AFB1-modified DNA (one AFB1 residue per 320 nucleotides) in unmodified calf thymus DNA. Wells contained 2.5 (0), 0.25 (0), 25 (Fl), or 5 (U) ng of AFB1-modified DNA.

476

PESTKA AND CHU

(2). Bassir and Bababunmi (2) found that HSA has two available binding sites for AFB1, whereas BSA has only one.

Our results suggest that at least one of the noncovalent H§A

binding sites for AFB1 may associate with or mask the cyclopentanone and methoxy end of the AFB1 molecule, whereas the single BSA binding site is specific for another part of the AFB1 molecule. However, since these albumin fractions were not completely purified, other factors, such as fatty acids, may be responsible for this interference. It is also significant that rabbit antisera prepared against AFB1-diol conjugate, as well as those prepared previously for AFB1 carboxymethyloxime (7, 16), AFM1 carboxymethyloxime (28), AFB2a hemiglutarate (17), and AFB2a (9, 29), showed cross-reactivity with AFB1 analogs which were dependent on the mechanism of conjugation to the protein carrier but that monoclonal antisera prepared against AFB1N7-Gua-modified DNA (10, 11) and AFB1 FaPyr-modified DNA (14) showed virtually no cross-reactivity with related analogs, such as AFB1-dibl, AFB2a, and AFB1-N7-Gua. Although this may be attributable to the high specificity of hybridoma antibodies, the latter two antibodies may have a high degree of recognition for alterations in the intact DNA molecule resulting from AFB1 modification rather than from the actual AF molecule. This possibility is supported by the fact that AFB1-FaPyr-modified DNA antibody is very reactive for both AFB1-modified IINA and AFG,-modified DNA (14). Since both AFB1 and AFG1 are bound to the guanine residue at C-2 of the AF molecule, the respective cyclopentanone and tetrahydropyranone rings should be most exposed. The fact that AFB1-FaPyr-modified DNA antibody cross-reacts with both these forms of modified DNA suggests that the primary epitope in the generation of this antibody was not the AFB1 molectule but rather some alteration in the modified DNA molectlle that is not present in unmodified DNA. In contrast, the cross-reactivity of AFB1-diol and AFB2a antisera with AFB1-modified DNA and closely related analogs suggests that part of the AF molecule was the primary epitope in their. generation. When a noncompetitive ELISA with an AFB1-modified DNA solid-phase immunosorbent was used, AFB, bound covalently to calf thymus DNA at a modification level of one AFB1 molecule per 200,000 nucleotides. Since this modification level has been shown to occur in experimental animals in vivo after exposure to AFB1, the noncompetitive ELISA might be using AFB1-diol antibody and might be sufficiently sensitive for monitoring covalent AFB, binding in vivo. The sensitivity of the assay is comparable to that described for assays with monoclonal antibodies prepared against AFB1FaPyr-modified DNA (one AFBi per 300,000 nucleotides [14]), but is lower than that of assays with monoclonal antibodies derived against AFB1-N7-Gua-modified DNA (one AFBI per 1,355,000 nucleotides [10]). The latter study employed ultrasensitive enzyme radioimmunoassay to attain high sensitivity. Thus, detection limits for AFBI-modified

DNA detection with AFB1-diol antibody might be similarly enhanced by using hybridoma technology and the ultrasensitive assay. High-performance liquid chromatography and radioisotopic methods are not completely adequate for monitoring AFB1 metabolism in vitro and in vivo. Specific antibodies against AFB, metabolites are useful alternatives to these approaches and might be used to discern the intracellular fate of AFB, in the liver and other organs at the ultrastructural level. We previously described the use of AFB2a antiserum in the immunoperoxidase light-microscopic localization of AFB, in livers of rats experimentally dosed with

APPL. ENVIRON. MICROBIOL.

this carcinogen (26). This study describes a simple procedure for the generation of AFB,-diol antiserum, the specificity of which was consistent with the proposed Schiff base mechanismn of AFB1-diol binding to primary amino groups. Thus, depending on the desired specificity, AFB1-diol, as well as AFB2a, antiserum can be used as a diagnostic tool for the detection of aflatoxicosis in humans and animals or as an immunological probe of AFB1 macromolecule interactions. ACKNOWLEDGMENTS This work was supported by the College of Agricultural and Life Sciehces, University of Wisconsin-Madison, by Public Health Service research grant CA15064 from the National Cancer Institute, and by Public Health Service training grant T32 ES07015 from the National Institute of Environmental Health Sciences. We thank Marlene Klaila for technical assistance with antibody preparation and Janeen Hunt for preparation of the manuscript. LITERATURE CITED 1. Ashoor, S. H., and F. S. Chu. 1975. Interaction of aflatoxin B2a with amino acids and proteins. Biochem. Pharmacol. 24: 1799-1805. 2. Bassir, O., and E. A. Bababunmi. 1973. The binding of aflatoxin B1 with albumin. Biochem. Pharmacol. 22:132-134. 3. Buchi, G. H,, K. C. Luk, and P. N. Muller. 1975. Synthesis of aflatoxin.Q1. J. Org. Chem, 40:3458-3460. 4. Bulatao-Jayme, J., E. M. Almero, M. C. Castro, M. T. Jardeleza, and L. A. Salamat. 1982. A case-control dietary study of primary liver cancer risk from aflatoxin exposure. Int. J. Epidemiol. 11:1-7. 5. Busby, W. F., and G. N. Wogan. 1979. Food-borne mycotoxins and alimentary mycotoxicoses, p. 519-610. In H. P. Riemann and F. L. Bryan (ed.), Food-borne infections and intoxications, 2nd ed. Academic Press, Inc., New York. 6. Chu, F. S. 1971. Chromatography of crude aflatoxin on adsorbsil-5. J. Assoc. Off. Anal. Chem. 54:1304-1306. 7. Chu, F. S., H. P. Lau, T. S. Fan, and G. S. Zhang. 1982. Ethylenediamine modified bovine serum albumin as protein carrier in the production of antibody against mycotoxins. J. Immunol. Methods 55:73-78. 8. Essigmann, J. M., R. G. Croy, A. M. Nadzan, W. F. Busby, and G. N. Wogan. 1977. Structural identification of the major DNA adduct formed by aflatoxin B, in vitro. Proc. Natl. Acad. Sci. U.S.A. 74:1870-1874. 9. Gaur, P. K., H. P. Lau, J. J. Pestka, and F. S. Chu. 1981. Production and characterization of aflatoxin B2a antiserum. Appl. Environ. Microbiol. 41:478-482. 10. Groopman, J. D., A. Haugen, G. R. Goodrich, G. N. Wogan, and C. C. Harris. 1982. Quantitation of aflatoxin B -modified DNA using monoclonal antibodies. Cancer Res. 42:3120-3124. 11. Haugen, A., J. D. Groopman, I. C. Hsu, G. R. Goodrich, G. N. Wogan, and C. C. Harris. 1981. Monoclonal antibody to aflatoxin B1-modified DNA detected by enzyme immunoassay. Proc. Natl. Acad. Sci. U.S.A. 18:4124-4127. 12. Hebert, G. A., P. L. Pelham, and B. Pittman. 1973. Determination of the optimal ammonium sulfate concentration for fractionation of rabbit, sheep, horse, and goat antisera. Appl. Microbiol. 25:26-36. 13. Hertzog, P. J., J. R. Lindsay Smith, and R. Colin Garner. 1982. Characterization of the imidazole ring-opened forms of trans8,9-di-hydro-8-(7-guanyl)9-hydroxy aflatoxin B,. Carcinogenesis (London) 3:723-725. 14. Het-tzog, P. J., J. R. Lindsay Smith, and R. Colin Garner. 1982. Production of monoclonal antibodies to guanine imidazole ringopened aflatoxin B, DNA, the persistent DNA adduct in vivo. Carcinogenesis (London) 3:825-828. 15. Hsu, I. C., M. C. Poirier, S. H. Yuspa, D. Grunberger, I. B. Weinstein, R. H. Yolken, and C. C. Harris. 1981. Measurement of benzo(a)pyrene-DNA adducts by enzyme immunoassays and radioimmunoassays. Cancer Res. 41:1091-1095. 16. Langone, J. L., and H. Van Vunakis. 1976. Aflatoxin BI-specific

VOL. 47, 1984

17.

18.

19.

20.

21. 22. 23.

24.

25. 26. 27.

antibodies and their use in radioimmunoassay. J. Natl. Cancer Inst. 56:591-595. Lau, H. P., P. K. Gaur, and F. S. Chu. 1981. Preparation and characterization of aflatoxin B2a-hemiglutarate and its use in production of antibody against aflatoxin Bl. J. Food Safety 3:113. Lin, J., K. A. Kennan, E. C. Miller, and J. A. Miller. 1978. Reduced nicotinamide adenine dinucleotide phosphate-dependent formation of 2,3-dihydro-2,3-dihydroxy-aflatoxin B1 from aflatoxin B1 by hepatic microsomes. Cancer Res. 38:2424-2428. Lin, J., J. A. Miller, and E. C. Miller. 1977. 2,3-Dihydro-2(guan-7-yl)-3-hydroxy-aflatoxin B1, a major acid hydrolysis product of aflatoxin B -DNA or ribosomal RNA adducts formed in hepatic microsome-mediated reactions and rat liver iin vivo. Cancer Res. 37:4430 4438. Martin, C. N., and R. C. Garner. 1977. Aflatoxin Bl-oxide generated by chemical or enzymic oxidation of aflatoxin B1 causes guanine substitution in nucleic acids. Nature (London) 267:863-865. Miller, E. C., and J. A. Miller. 1979. Naturally occurring chemical carcinogens that may be present in foods. Int. Rev. Biochem. 27:123-165. Neal, G. E., and P. J. Colley. 1979. The formation of 2,3dihydro-2,3-dihydroxy aflatoxin B1 in vitro by rat liver microsomes. FEBS Lett. 101:382-386. Neal, G. E., D. J. Judah, F. Stripe, and D. S. Patterson. 1981. The formation of 2,3-dihydroxy-2,3-dihydro-aflatoxin B1 by the metabolism of aflatoxin B, by liver microsomes isolated from certain avian and mammalian species and the possible role of this metabolite in the acute toxicity of aflatoxin Bl. Toxicol. Appl. Pharmacol. 58:431-434. Oellerich, M. 1980. Enzyme immunoassays in clinical chemistry: present status and trends. J. Clin. Chem. Clin. Biochem. 18:197-208. Peers, F. G., and C. A. Linsell. 1973. Dietary aflatoxins and liver cancer-a population based study in Kenya. Br. J. Cancer 27:473-481. Pestka, J. J., J. T. Beery, and F. S. Chu. 1983. Indirect immunoperoxidase localization of aflatoxin B1 in rat liver. Food. Chem. Toxicol. 21;41-48. Pestka, J. J., P. K. Gaur, and F. S. Chu. 1980. Quantitation of aflatoxin B1 and aflatoxin B1 antibody by an enzyme-linked immunosorbent microassay. Appl. Environ. Microbiol.

40:1027-1031. 28. Pestka, J. J., Y. K. Li, and F. S. Chu. 1982. Reactivity of

AFB,-DIOL ANTIBODY

29.

30. 31.

32.

477

aflatoxin B2a antibody with aflatoxin Bl-modified DNA and related metabolites. Appl. Env. Microbiol. 44:1159-1165. Pestka, J. J., Y. Li, W. 0. Harder, and F. S. Chu. 1981. Comparison of radioimmunoassay and enzyme-linked immunosorbent assay for determining aflatoxin Ml in milk. J. Assoc. Off. Anal. Chem. 64:294-301. Peters, H., M. P. Dierich, and K. Dose. 1982. Enzyme-linked immunosorbent assay for detection of T-2 toxin. HoppeSeyler's Z. Physiol. Chem. 363:1437-1441. Pier, A. C., J. L. Richard, and S. S. Cysewski. 1980. Implications of mycotoxins in animal disease. J. Am. Vet. Med. Assoc. 8:719-724. Pohland, A. E., M. E. Cushman, and P. J. Andrellos. 1968. Aflatoxin B1 hemiacetal. J. Assoc. Off. Anal. Chem. 51:907-

910. 33. Rao, V. N. U., K. Valmikinathan, and N. Verghese. 1968. Binding of aflatoxin by plasma albumin in vitro. Biochim. Biophys. Acta 165:288-289. 34. Scoppa, P., and W. 0. Borle. 1971. Interazione dell'aflatossina B1 con 1' albumin: studio cromatografico mediante il metodo dell'eluzione frontale. Boll. Soc. Ital. Biol. Sper. 47:201-204. 35. Scoppa, P., and E. Marafante. 1971. Interazione dell'aflatossina B1 con l'albumina: studio spetrofotometrico mediante esame degli spettri differenziali. Boll. Soc. Ital. Biol. Sper. 47:198-200. 36. Shank, R. C. 1978. Mycotoxicoses of man: dietary and epidemiological conditions, p. 1-19. In T. D. Wylie and L. G. Morehouse (ed.), Mycotoxic fungi, mycotoxins, mycotoxicoses: an encyclopedic handbook, vol. 3. Marcel Dekker, Inc., New York. 37. Shank, R. C., N. Bhamarapravati, J. E. Gordon, and G. N. Wogan. 1972. Dietary aflatoxins and human liver cancer. IV. Incidence of primary liver cancer in two municipal populations of Thailand. Food Cosmet. Toxicol. 10:171-179. 38. Stark, A. A. 1980. Mutagenicity and carcinogenicity of mycotoxins: DNA binding as a possible mode of action. Annu. Rev. Microbiol. 34:235-262. 39. Swenson, D. H., J. A. Miller, and E. C. Miller. 1975. The reactivity and carcinogenicity of aflatoxin B1-2,3 dichloride, a model for the putative 2,3-oxide metabolite of aflatoxin B1. Cancer Res. 55:3811-3823. 40. Wang, T. V., and P. A. Cerutti. 1980. Spontaneous reactions of aflatoxin B1 modified deoxyribonucleic acid in vitro. Biochemistry 19:1692-1698. 41. Wei, R. D., and S. S. Lee. 1971. Binding of aflatoxins B1 and G, to human serum proteins. Experientia 27:458-460.