Site-Specific Synthesis of Aflatoxin B1 Adducts within an

Wendelyn R. Jones, David S. Johnston, and Michael P. Stone*. Department of Chemistry and Center in Molecular Toxicology, Vanderbilt University, Nashvi...
0 downloads 0 Views 105KB Size
Download PDF
Chem. Res. Toxicol. 1999, 12, 707-714

707

Site-Specific Synthesis of Aflatoxin B1 Adducts within an Oligodeoxyribonucleotide Containing the Human p53 Codon 249 Sequence Wendelyn R. Jones,† David S. Johnston, and Michael P. Stone* Department of Chemistry and Center in Molecular Toxicology, Vanderbilt University, Nashville, Tennessee 37235 Received March 22, 1999

This work describes the preparation of the cationic trans-8,9-dihydro-8-(N7-guanyl)-9hydroxyaflatoxin B1 (AFBG) adducts at the positions corresponding to G746 or G747, within the oligodeoxyribonucleotide d(GGAGGCCT) containing the codon 249 sequence (underlined) of the p53 gene, using DNA triplexes to target adduction at the desired site. This approach enabled the successful preparation and purification of sufficient quantities of d(GGAGAFBGCCT) for NMR structural studies, using only standard phosphoramidites. The presence of multiple guanines in this oligodeoxynucleotide precluded the direct reaction of d(GGAGGCCT)‚ d(AGGCCTCC) with aflatoxin epoxide as a method for producing large quantities of site-specific adducts for physical studies. Of the multiple potential alkylation sites at guanine N7 in d(GGAGGCCT)‚d(AGGCCTCC), it was found that sites G2 and G5 exhibited approximately equal reactivity with aflatoxin B1-exo-8,9-epoxide; the reactivity at site G4 was reduced by approximately a factor of 2 as compared to that at G2 or G5. To successfully prepare the sitespecific adducts, the p53 oligodeoxyribonucleotide was annealed with either the blocking strand d(CTCCATTTTCCT) or d(CCTCCATTTTCCTC) to form the corresponding partial triplexes which targeted AFB1 adduction either to G4 or to G5. Piperidine cleavage, followed by heating, confirmed that in each instance, the product corresponded to the lone guanine not protected from adduction by the partial DNA triplex. The adducted oligodeoxyribonucleotides were examined with regard to purity by capillary electrophoresis. The primary advantage of this modified triple helix methodology is that it requires only standard phosphoramidites; thus, it is applicable to large-scale preparations that are necessary for NMR structural studies or other physical measurements.

Introduction 1

Aflatoxin B1 (AFB1) is the predominant mutagenic fungal metabolite isolated from several species of Aspergillus. This mycotoxin contaminates many food products and is consequently of worldwide health concern. AFB1 is a mutagen in several strains of bacteria (1), and a hepatocarcinogen in animals (2, 3). Epidemiological studies suggest it may be a carcinogen in humans (2, 4). Furthermore, AFB1 may be linked to site-specific transversion in the tumor suppressor gene p53 (5, 6) and to protooncogene activation (3, 7). AFB1 is metabolized by cytochrome P450 monooxygenases (8, 9) to yield the ultimate carcinogen, AFB1-exo8,9-epoxide (10, 11). The exo-epoxide bonds predominantly to guanine N7 to yield trans-8,9-dihydro-8-(N7guanyl)-9-hydroxyaflatoxin B1 (12); small amounts of the corresponding adenine adduct trans-8,9-dihydro-8-(N7adenyl)-9-hydroxyaflatoxin B1 also can form (Scheme 1) * To whom correspondence should be addressed. Telephone: (615) 322-2589. Fax: (615) 343-1234. E-mail: [email protected]. edu. † Present address: Division of Bioengineering and Environmental Health, 56-787, Massachusetts Institute of Technology, Cambridge, MA 02139-4307. 1 Abbreviations: AFB , aflatoxin B ; CGE, capillary gel electro1 1 phoresis; EDTA, ethylenediaminetetraacetic acid; PAGE, polyacrylamide gel electrophoresis; PRP, polymeric reversed phase.

(13). AFB1-exo-8,9-epoxide exhibited strong regioselectivity for the N7 position of guanine (12, 14); minor amounts of adduction at adenine N7 can also be observed (13, 15). The synthesis of AFB1-exo-8,9-epoxide allowed sufficient quantities of site-specific AFB1 adducts to be produced for NMR structural and related biophysical studies (16-20), and for site-directed mutagenesis studies (21, 22). Structural studies of the related sterigmatocystin adduct have also been reported (23, 24). Many DNA sequences of biological interest, including that of the p53 tumor suppressor gene, contain multiple guanines. Reports linking AFB1 to site-specific transversion in the tumor suppressor gene p53 (5, 6, 25-29), possibly exacerbated by co-infection by the hepatitis B virus (28, 30-32), made facile routes to these adducts important. The nonspecific reaction with AFB1-exo-8,9epoxide (10) was limited by the inability to control sequence selectivity in oligodeoxyribonucleotides containing multiple guanines. Nonspecific reaction with the p53 sequence was expected to result in multiple adducts, at low yield, necessitating complex separation protocols for isolating the desired product. The AFB1-exo-8,9-epoxide reaction was targeted to a specific site by selectively incorporating a covalent 5′intercalating agent, the cis-syn thymidine benzofuran (33, 34), opposite nontargeted guanines in an oligodeoxyri-

10.1021/tx990048u CCC: $18.00 © 1999 American Chemical Society Published on Web 07/16/1999

708 Chem. Res. Toxicol., Vol. 12, No. 8, 1999

Jones et al.

Scheme 1. Formation of the Primary AFB1-Guanine N7 Adduct from the Reaction of AFB1-exo-Epoxide with DNAa

a The epoxide was prepared synthetically by reaction of AFB with dimethyldioxirane; the epoxide is enzymatically produced in vivo by 1 cytochrome P450s.

bonucleotide duplex containing the p53 mutational hotspot (35). The adduction at guanine N7 was mechanistically consistent with precovalent intercalation of the epoxide on the 5′ face of the guanine (16, 36-39) prior to SN2 attack by the epoxide (40). The cis-syn thymidine benzofuran approach (33, 34) would appear to have two potential disadvantages for the preparation of the large quantities of site-specific adducts required for structural studies. The first is the need for stoichiometric quantities of a blocking strand containing an exotic phosphoramidite. Second, incorporation of the covalently linked intercalator might be expected to result in nearest-neighbor exclusion effects at adjacent binding sites. Accordingly, it seemed advantageous to develop a synthetic route requiring only standard phosphoramidites, which could be successfully applied to large-scale syntheses. The observation that the third strand of a DNA triplex could target site-specific cleavage of DNA (41-44) suggested an alternative route to large-scale sitespecific incorporation of AFB1 in the p53 gene. This was demonstrated by the ability of nicked DNA triplexes to site-specifically direct AFB1 adduction within an iterated repeat of five guanines (45). In that instance, the third strand of DNA in the major groove blocked reactivity at the N7 positions of guanines involved in triplex formation, allowing adduction to be targeted in a site-specific manner to nonprotected guanines. The selectivity of the nicked triplex approach was dependent upon equilibrium binding between the triplex-forming blocking strand and the targeted duplex. This was not a problem in the smallscale syntheses used in the previous study (45). However, there were concerns about whether this approach could be successfully scaled to produce larger quantities of sitespecific adducts required for NMR analysis. Furthermore, to utilize the triple helix approach for site-specifically positioning AFB1 adducts in the p53 sequence, modification of the chemistry was necessitated because the p53 sequence was not a homopurine sequence, and the

targeted guanine was flanked on the 3′ side by two cytosines. Using canonical triplets, it was not possible to design a blocking strand that would adopt a complete nicked triplex with the targeted strand. One potentially could exploit noncanonical triplets (46-48), but similar to the cis-syn thymidine benzofuran approach, this was judged to be undesirable for large-scale preparations. Moreover, the 3′-flanking cytosines created a problem because during targeted adduction of AFB1-exo-8,9-epoxide, the epoxide could also react with the blocking strand at the guanines complementary to the 3′-flanking cytosines. The resulting structural perturbations introduced by the presence of the adventitious adducted guanines disrupt blocking strand specificity for the targeted guanine. This work demonstrates the applicability of a modified version of the nicked triplex approach (45) to the sitespecific preparation of AFB1 adducts in the p53 sequence. In the modified approach, the p53 oligodeoxyribonucleotide d(GGAGGCCT), containing the codon 249 sequence (underlined), including G746 and G747 (bold), was annealed with blocking strands designed to form only partial triplexes, leaving the 3′-flanking cytosines unprotected. Adduction was targeted to G4, at the position corresponding to G746 of the p53 gene, with the blocking strand d(CTCCATTTTCCT). Adduction was targeted to G5, at the position corresponding to G747 of the p53 gene, with the blocking strand d(CCTCCATTTTCCTC). The ability to prepare large quantities of site-specific aflatoxin adducts in the p53 oligodeoxyribonucleotide will enable structural studies to be conducted.

Experimental Procedures Adduct Synthesis. Unadducted oligodeoxyribonucleotides were purchased from Midland Certified Reagent Co. (Midland, TX), or synthesized in-house. AFB1 was purchased from SigmaAldrich Chemicals, Inc. (Milwaukee, WI). Dimethyldioxirane was reacted with AFB1 to give AFB1-exo-8,9-epoxide (10).

Aflatoxin B1 Adduct in the p53 Codon 249 Sequence Caution: Crystalline aflatoxins are hazardous due to their electrostatic nature and should be handled using appropriate containment procedures and a respiratory mask to prevent inhalation. Aflatoxins can be destroyed by treatment with NaOCl. It should be assumed that aflatoxin epoxides are highly toxic and carcinogenic. Manipulations should be carried out in a well-ventilated hood with suitable containment procedures. The epoxide was dissolved at a concentration of 30 mM in methylene chloride. To form the p53 adducts, two 250 µL aliquots of the epoxide were added sequentially to 30 nmol of the triplestranded oligodeoxyribonucleotide complexes in 200 µL of TH buffer [10 mM Na2HPO4, 100 mM NaCl, 10 mM MgCl2, and 0.05 mM Na2EDTA (pH 5.8)] to form a two-phase mixture. The mixtures were vortexed and reacted for 15 min at 5 °C, with a 5 min interval between additions. The molar ratio of epoxide to oligodeoxyribonucleotide was 0.5:1. Gel Electrophoresis. The oligodeoxyribonucleotides were labeled with 32P using T4 polynucleotide kinase (Promega Corp.) and [γ-32P]ATP (NEN/Dupont, Newark, DE). Unincorporated nucleotide was removed using spin columns packed with Sephadex G-25 (super fine mesh, Bio-Rad, Hercules, CA). The eluent was reacted with dimethyl sulfate for G-specific cleavage, with formic acid for G- and A-specific cleavage, or with AFB1-exo8,9-epoxide to generate base labile sites (15, 49, 50). β-Elimination of the phosphate backbone was accomplished by addition of piperidine, followed by heating. The samples were lyophilized. Loading buffer (Amersham Pharmacia Biotech, Piscataway, NJ) was added to the dried aliquots prior to loading the samples onto the denaturing polyacrylamide gel. A 20% (19:1 acryl:bis) (Bio-Rad) 7 M urea, 0.4 mm gel was used for electrophoretic separation. The gel was exposed to X-OMAT film to generate the autoradiogram. The autoradiograms were analyzed densitometrically with a Macintosh computer using the public domain NIH Image program (developed at the National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/ nih-image/). Adduct Purification. The modified oligodeoxyribonucleotides destined for NMR studies were maintained at 5 °C during purification by HPLC using a reverse-phase semipreparative column (PRP-1, Hamilton Co., Reno, NV; or an Econosil C-18, Alltech, Chicago, IL) equilibrated with 5 mM sodium phosphate buffer (pH 7.5). The critical aspects of the purification were a slightly basic pH of 7.5 and the low temperature, which minimized the opening of the imidazole ring of the modified sample. The oligodeoxyribonucleotides were eluted using a gradient consisting of 0 to 30% acetonitrile over the course of 40 min. The products were further examined with regard to purity by capillary gel electrophoresis using a Beckman PACE 5000 instrument and an oligodeoxyribonucleotide CGE kit from Beckman Instruments. The sample sizes were approximately 0.075 OD unit (254 nm) dissolved in 100 µL of water. The samples were loaded at 10 kV for 2 s. The separation voltage of 10 kV was maintained for 25 min. The gel was maintained at 30 °C. The electropherograms were monitored at 254 nm. Sequence Specificity of the Aflatoxin Reaction in the p53 Oligomer. To test the sequence-specific reactivity differences of the four guanines in the targeted p53 strand, the p53 oligodeoxyribonucleotide was 32P end-labeled and annealed to its complementary strand d(AGGCCTCC). The resulting duplex was reacted with AFB1-exo-8,9-epoxide at a molar ratio of 0.5:1 epoxide:DNA. This reaction mixture was treated with piperidine to generate backbone cleavage at the site of the lesion. The ratio of epoxide:DNA was controlled to limit the formation of adducts to one adduct per duplex. The resultant autoradiogram was densitometrically analyzed by NIH image as described above.

Results and Discussion The current understanding of the mechanism for AFB1 adduction at guanine N7 suggests precovalent intercalation of the epoxide initially occurs above the 5′ face of guanine in duplex DNA, followed by SN2 attack by guanine N7 on the intercalated epoxide. Several lines of

Chem. Res. Toxicol., Vol. 12, No. 8, 1999 709 Scheme 2. Strategy for the Successful Targeting of AFB1-8,9-Epoxide at G4 and G5 of the p53 8-mer Oligodeoxyribonucleotide Using the Partial Triple Helix Approach

evidence support this mechanism (51). Aflatoxin G1, which substitutes a δ-lactone ring for the cyclopentenone ring of aflatoxin B1, exhibits a lower affinity for binding to B-DNA, and forms lower levels of DNA adducts (38). AFB1-exo-8,9-epoxide reacts readily with B-form DNA, but not with A- or Z-form DNA, or single-stranded DNA (39). Reaction with the sequence isomeric oligodeoxyribonucleotides d(ATCGAT)2 and d(ATGCAT)2 proceeds with differing stoichiometries. d(ATCGAT)2 will react with only 1 equiv of epoxide, while d(ATGCAT)2 will react with 2 equiv (19). AFB1-endo-8,9-epoxide (52) does not form DNA adducts and is not mutagenic (40). The epoxide is highly reactive in aqueous solution, and hydrolyzes to the dihydrodiol with a rate constant of 0.6 s-1 (53), which probably accounts for the low yield of adducts in singlestranded DNA and mononucleotides. Previous structural studies with AFB1 and structurally related adducts (17-20, 23, 24) utilized duplexes containing either single guanines or multiple guanines which were symmetrical about the pseudodyad axis of the DNA. This avoided problems associated with the isolation of a single adduct from a mixture of adducts at different guanines in a duplex containing multiple guanines. On the other hand, DNA sequences of mutagenic significance often contain multiple guanines. In the 8-mer oligodeoxynucleotide duplex containing the codon 249 sequence of p53 chosen for this study, the targeted guanines G4 and G5 were surrounded by four additional G‚C base pairs (Scheme 2). The reaction of duplex DNA with AFB1-epoxide was anticipated to result in a complex mixture of products. The relative reactivity of AFB1-exo-8,9-epoxide with differing guanines in the template strand of the duplex p53 oligodeoxyribonucleotide was examined. The duplex oligodeoxyribonucleotide was reacted with the epoxide under single-hit conditions. Following piperidine-induced cleavage of the 32P-5′-endlabeled oligodeoxynucleotide and denaturing PAGE, the intensity of the cleavage band at each of the three guanines was monitored by densitometric analysis (data not shown). The experiment was repeated in triplicate. Intensities were normalized relative to the intensity of the G5 cleavage band. In the p53 duplex, both G2 and G5 exhibited similar reactivities with respect to AFB1-exo8,9-epoxide, while the reactivity at G4 was lower by a factor of approximately 0.5. These results led to the conclusion that attempts to isolate and purify large quantities of site-specific adducts at G4 or G5 from a mixture of adducts produced by random adduction of the p53 duplex with AFB1 epoxide would be futile. This difficulty previously led to the clever

710 Chem. Res. Toxicol., Vol. 12, No. 8, 1999

design of a method by which the AFB1-exo-8,9-epoxide reaction was targeted to a specific site by selectively incorporating a covalent 5′-intercalating agent, the cissyn thymidine benzofuran (33, 34), opposite nontargeted guanines (35). The random adduction experiments conducted under single-hit conditions were consistent with the notion that guanine N7 alkylation was modulated by sequence context, with the identity of the 5′ neighbor nucleotide being a major contributor (54, 55). There has been considerable interest in adduction and mutagenesis at CpG sites (56, 57), which are sites of methylation (58). In codon 249 of p53, neither G4 nor G5 was a CpG site. While G2 and G5 had a 5′ neighbor guanine, G4 had a 5′ neighbor adenine. Interestingly, the G4 position was also the one purine f pyrimidine base step in this duplex. A displacement of the purine toward the minor groove was associated with purine f pyrimidine steps, proposed to relieve steric clashing of the purines (59, 60). We conclude that G5 represents a site that exhibits reasonable reactivity with AFB1-exo-8,9-epoxide. To the extent that AFB1 adduction at this site results in activating mutations of the p53 tumor suppressor gene, the tendency toward these mutations may be exacerbated by the chemical reactivity at this site. Development of a Modified Triplex Methodology. The initial concern was whether the triplex DNA approach (45) could be successfully modified to be applied to the preparation of site-specific adducts in the p53 oligodeoxynucleotide. The p53 oligomer was not a homopurine sequence. There was no means by which to design a blocking strand that would adopt a complete nicked triplex (45) with the targeted strand. In one experiment, d(CCTTGTTTTAAGGCCTCCATTTTCCT) was designed to target G4, and d(CTTGTTTTAAGGCCTCCATTTTCCTC) was designed to target G5. These blocking strands caused cytosines C6 and C7 to be paired with thymine in the blocking strand. Polyacrylamide gel electrophoresis analysis of the piperidine-induced cleavage products revealed a lack of site selectivity. The use of noncanonical base triplets (46-48) was considered. However, this would have required a synthetically intensive and potentially costly preparation of the blocking strand. The cytosines flanking the targeted guanines in the 3′ direction created an additional problem because during targeted adduction of AFB1-exo-8,9-epoxide, the epoxide could also react with the triplex blocking strand at the guanines complementary to the 3′-flanking cytosines. The resulting perturbations introduced by the presence of the adventitious adducted guanines disrupted blocking strand specificity for the targeted guanine. After several iterations, both to the targeted strand and to the blocking strand, it was found that a modification of the triplex methodology, in which the blocking strands did not fully wrap around the targeted strand, would work. Scheme 2 shows in detail the successful methodology for targeting G5 in the targeted strand d(GGAGGCCT). The 14-mer blocking strand d(CCTCCATTTTCCTC) was utilized. While the first nucleotides from the 5′ end of the targeted strand formed triplets with the blocking strand, the trailing three pyrimidines 5′-‚‚‚CCT3′ at the 3′ end of the target strand remained singlestranded. The targeted guanine, G5, formed a WatsonCrick base pair with the complementary cytosine in the blocking strand. This Watson-Crick pair at G5 was probably susceptible to fraying; however, it fulfilled the

Jones et al.

Figure 1. 32P autoradiogram of piperidine cleavage products following reaction of the 8-mer p53 bimolecular triplex to target G5 (corresponding to G747 in the p53 gene sequence) with AFB1exo-8,9-epoxide: lane 1, probe lane; lane 2, guanine-specific reaction; lane 3, guanine- and adenine-specific reaction; and lane 4, aflatoxin reaction.

general criterion for double-stranded DNA at the targeted guanine (40). This design was advantageous in that it did not require addition of guanine residues opposite the two cytosines that were in the 3′ direction from the targeted guanine. Accordingly, it eliminated the problem of adventitious reaction with the added guanine residues. The 14-mer was annealed to the 32P-labeled 8-mer. The triplex was reacted with AFB1-exo-8,9-epoxide. The labeled 8-mer was subjected to piperidine-induced backbone cleavage (15). DNA sequencing (49) was used to identify the lesion site at G5 (Figure 1). The partial triplex methodology was extended to the site-specific adduction at G4 in d(GGAGGCCT). In this instance, the blocking strand d(CTCCATTTTCCT) targeted G4. The 12-mer blocking strand was annealed to the 32P-labeled 8-mer, resulting in spontaneous formation of the triplex shown in Scheme 2. The triplex was reacted with AFB1-exo-8,9-epoxide. Following piperidine-induced backbone cleavage of the oligomer and DNA sequencing, analysis by polyacrylamide gel electrophoresis indicated that the d(CTCCATTTTCCT) blocking strand targeted G4 (Figure 2). The successful site-specific adduction of G4 illustrated the preference for aflatoxin adduction in double helices (39). In this case, the targeted G4 residue was involved in a Watson-Crick base pair with a cytosine, while the nontargeted G5 position was singlestranded. The other two guanines, G1 and G2, were involved in CGC base triplets. Thus, in this triple helix G4 only reacted readily with the epoxide. The presence of the 3′ terminal thymine on the targeted p53 oligomer (Scheme 2) proved to be crucial to this particular modification of the DNA triplex methodology. The initial codon 249-containing oligodeoxyribonucleotide sequence chosen for the partial triplex modification was a 7-mer that lacked the 3′-terminal thymine, d(GGAGGCC). In that instance (Figure 3), adduction at G5, the targeted guanine, was observed. However, a secondary adduct was observed at G2. The observation of a secondary reaction product at G2 but not at G4 using the 7-mer target strand was unexpected since G2 was expected to be protected by the triplex blocking strand

Aflatoxin B1 Adduct in the p53 Codon 249 Sequence

Chem. Res. Toxicol., Vol. 12, No. 8, 1999 711 Scheme 3. Essential Presence of the Terminal 3′-Thymine on the Targeted Stranda

Figure 2. 32P autoradiogram of piperidine cleavage products following reaction of the 8-mer p53 bimolecular triplex to target G4 (corresponding to G746 in the p53 gene sequence) with AFB1exo-8,9-epoxide: lane 1, probe lane; lane 2, guanine-specific reaction; lane 3, guanine- and adenine-specific reaction; and lane 4, aflatoxin reaction.

a In the absence of the 3′-thymine, simultaneous adduct formation at both G2 (minor product) and G5 (major product) using the indicated 7-mer as the target strand occurred as a consequence of equilibration between 5′- and 3′-hairpin loop formation involving the blocking 14-mer strand.

Figure 4. UV/vis spectrum of the site-specifically modified oligodeoxyribonucleotide d(GGAGAFBGCCT), targeting G5 (corresponding to G747 in the p53 gene sequence).

Figure 3. 32P autoradiogram of piperidine cleavage products following reaction of the 7-mer p53 bimolecular triplex to target G747 with AFB1-exo-8,9-epoxide: lane 1, probe; lane 2, guaninespecific reaction; lane 3, guanine- and adenine-specific reaction; and lane 4, aflatoxin reaction.

as shown in Scheme 2. The minor product was most likely a consequence of the formation of a hairpin loop about the 3′ terminus instead of the 5′ terminus of the p53 oligomer (Scheme 3). Addition of a thymine to the 3′ terminus of the targeted strand in d(GGAGGCCT) (Scheme 2) presumably lowered the stability of the unwanted DNA triplex which targeted G2, thus selectively favoring the desired triplex which targeted G5. In this instance, the choice of adding thymidine as opposed to some other nucleotide at the 3′ terminus of the targeted oligomer was dictated by spectroscopic expediency. An additional thymidine was anticipated to simplify the NMR spectrum. Thymidine was also the least likely nucleotide to bind unexpectedly to the blocking strand, and inadvertently stabilize the unwanted hairpin loop about the 3′ terminus of the p53 oligodeoxynucleotide. Purification of Site-Specifically Modified Codon 249 Adducts. Following reaction with AFB1-epoxide, the

reaction mixtures containing the site-specifically modified 8-mers were separated by reverse phase HPLC. The following discussion details the purification of the G5 adduct; a similar methodology applies to the G4 adduct. In the case of the G5 adduct, a peak at 16 min exhibited absorbance at 260 nm. It was identified as the unadducted target strand. A peak at 20 min was identified as the adducted strand d(GGAGAFBGCCT); it exhibited absorption at 260 and 360 nm. It was obtained in approximately 40% yield from the unmodified target strand. A large peak, which eluted at 23 min, contained the unadducted blocking strand. The desired adducted oligomer d(GGAGAFBGCCT) was subjected to UV spectroscopic analysis. The absorbance at 265 nm was ∼8 times greater than the absorbance at 365 nm (Figure 4). The purified sample of d(GGAGAFBGCCT) was analyzed by capillary gel electrophoresis (CGE). The electropherogram is shown in Figure 5. A single peak was observed. During adduct purification, the pH was increased relative to the reaction conditions, which were at pH 5.8, to minimize depurination of the cationic adduct during the chromatographic separation and collection. The potential for base-catalyzed hydrolysis occurring at C8 leading to formation of the trans-8,9-dihydro-8-(2,6diamino-4-oxo-3,4-dihydropyrimid-5-ylformamido)-9-hydroxyaflatoxin B1 (FAPY) adduct, due to the fact that the

712 Chem. Res. Toxicol., Vol. 12, No. 8, 1999

Jones et al.

Figure 5. Capillary gel electrophoresis of the site-specifically modified oligodeoxyribonucleotide d(GGAGAFBGCCT), targeting G5 (corresponding to G747 in the p53 gene sequence).

HPLC separation occurred at slightly alkaline pH, was not a serious problem. FAPY formation was slow compared to the time scale of purification procedures at slightly alkaline pH (7-8.5). Even at pH 9 and 50 °C, FAPY formation remained slow (on the order of hours for complete conversion to FAPY). At mildly alkaline pH, the cationic adduct was sufficiently stable over the time course of purification. Manipulation of the cationic adducts between pH 7.5 and 8.0 minimized FAPY formation. Nevertheless, small amounts of FAPY adducts which could arise during the purification process might pose a problem in purifying material for mutagenesis or other studies where absolute purity is crucial. Large-Scale Synthesis for Physical Studies. The second concern was whether the modified DNA triplex approach could be successfully scaled up to produce the larger quantities of site-specific adducts required for physical studies. To investigate this, the site-specific adduct d(GGAGAFBGCCT) at G5, corresponding to G747 of the p53 sequence, was prepared on a larger scale to obtain sufficient quantities for NMR structural studies. The products of four separate syntheses were collected and combined. Approximately 30 OD units of material was prepared in each synthesis. The difference in HPLC retention time of the adducted 8-mer target strand compared to that of the longer 14-mer blocking strand was relatively short. This was probably a consequence of the high purine content and the presence of the adduct in the targeted 8-mer, as compared to the high pyrimidine content of the blocking strand. In preparative-scale separations, two separate passes on the HPLC were performed to ensure adequate purification of the adducted 8-mer. The first pass allowed complete separation of the unadducted 8-mer from the reaction mixture as well as partial isolation of the blocking strand and secondary adducts. A second HPLC separation maximized the yield of adduct which was sufficiently pure for NMR spectroscopy. A 1H NMR spectrum of the adduct is shown in Figure 6. The AFB1 H6a proton resonance was identified at 6.7 ppm, from the scalar coupling to the H8, H9, and H9a protons of the AFB1 terminal furan ring. This provided a useful diagnostic resonance in 1H NMR spectra, which allowed identification of the adducted oligomer. The H6a proton characteristically resonates upfield from the DNA aromatic protons and downfield from the DNA anomeric protons (16). Integration confirmed a 1:1 ratio of the intensities of the H6a resonance to the individual nucleotide base aromatic resonances of the oligodeoxyribonucleotide. NMR spectroscopy places less rigorous demands upon sample purity than does site-

Figure 6. Preparation of NMR quantities of d(GGAGAFBGCCT), targeting G5 (corresponding to G747 in the p53 gene sequence). 1H NMR data were recorded at 10 °C.

specific mutagenesis in vivo (21, 22, 61). Thus, while the large-scale sample was successfully prepared in sufficient quantity for NMR applications, more rigorous purification might be necessary if the sample were to be applied to mutagenesis experiments. Summary. The primary advantage of this modified triple-helix methodology is that it requires only standard phosphoramidites; thus, it can readily be applied to largescale preparations that are necessary for NMR structural studies or other physical measurements. The approach was successfully applied to target large-scale AFB1 adduction in codon 249 of the p53 tumor suppressor gene. The resulting site-specific adducts will be the focus of future NMR structural studies. These future studies may also provide information about whether the differential chemical reactivities between the G4 and G5 positions in the p53 oligomer have a structural basis.

Acknowledgment. This work was supported by NIH Grant CA-55678 (M.P.S.) and the Molecular Genetics Core Facilities of the Vanderbilt Center in Molecular Toxicology (ES-00267). W.R.J. acknowledges support from a training grant in Molecular Biophysics (GM08320).

References (1) McCann, J., Spingarn, N. E., Koburi, J., and Ames, B. N. (1975) Detection of carcinogens as mutagenssbacterial tester strains with R-factor plasmids. Proc. Natl. Acad. Sci. U.S.A. 72, 979983. (2) Busby, W. F., Jr., and Wogan, G. N. (1984) Aflatoxins. In Chemical Carcinogens (Searle, C. E., Ed.) pp 945-1136, American Chemical Society, Washington, DC. (3) McMahon, G., Davis, E. F., Huber, L. J., Kim, Y., and Wogan, G. N. (1990) Characterization of c-Ki-ras and N-ras oncogenes in aflatoxin B1-induced rat liver tumors. Proc. Natl. Acad. Sci. U.S.A. 87, 1104-1108. (4) Wogan, G. N. (1992) Aflatoxin carcinogenesis: Interspecies potency differences and relevance for human risk assessment. Prog. Clin. Biol. Res. 374, 123-137. (5) Hsu, I. C., Metcalf, R. A., Sun, T., Welsh, J. A., Wang, N. J., and Harris, C. C. (1991) Mutational hotspot in the p53 gene in human hepatocellular carcinomas. Nature 350, 427-428. (6) Bressac, B., Kew, M., Wands, J., and Ozturk, M. (1991) Selective G to T mutations of p53 in hepatocellular carcinoma from southern Africa. Nature 350, 429-431. (7) Soman, N. R., and Wogan, G. N. (1993) Activation of the c-Ki-ras oncogene in aflatoxin B1-induced hepatocellular carcinoma and

Aflatoxin B1 Adduct in the p53 Codon 249 Sequence

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

adenoma in the rat: Detection by denaturing gradient gell electrophoresis. Proc. Natl. Acad. Sci. U.S.A. 90, 2045-2049. Shimada, T., and Guengerich, F. P. (1989) Evidence for cytochrome P-450NF, the nifedipine oxidase, being the principal enzyme involved in the bioactivation of aflatoxins in human liver. Proc. Natl. Acad. Sci. U.S.A. 86, 462-465. Aoyama, T., Yamano, S., Guzelian, P. S., Gelboin, H. V., and Gonzalez, F. J. (1990) Five of 12 forms of vaccinia virus-expressed human hepatic cytochrome P450 metabolically activate aflatoxin B1. Proc. Natl. Acad. Sci. U.S.A. 87, 4790-4793. Baertschi, S. W., Raney, K. D., Stone, M. P., and Harris, T. M. (1988) Preparation of the 8,9-epoxide of the mycotoxin aflatoxin B1: The ultimate carcinogenic species. J. Am. Chem. Soc. 110, 7929-7931. Iyer, R., and Harris, T. M. (1993) Preparation of Aflatoxin B1 8,9epoxide using m-chloroperbenzoic acid. Chem. Res. Toxicol. 6, 313-316. Essigmann, J. M., Croy, R. G., Nadzan, A. M., Busby, W. F., Jr., Reinhold, V. N., Buchi, G., and Wogan, G. N. (1977) Structural identification of the major DNA adduct formed by aflatoxin B1 in vitro. Proc. Natl. Acad. Sci. U.S.A. 74, 1870-1874. Iyer, R. S., Voehler, M. W., and Harris, T. M. (1994) Adenine adduct of aflatoxin B1 epoxide. J. Am. Chem. Soc. 116, 88638869. Croy, R. G., Essigmann, J. M., Reinhold, V. N., and Wogan, G. N. (1978) Identification of the principal aflatoxin B1-DNA adduct formed in vivo in rat liver. Proc. Natl. Acad. Sci. U.S.A. 75, 17451749. D’Andrea, A. D., and Haseltine, W. A. (1978) Modification of DNA by aflatoxin B1 creates alkali-labile lesions in DNA at positions of guanine and adenine. Proc. Natl. Acad. Sci. U.S.A. 75, 41204124. Gopalakrishnan, S., Stone, M. P., and Harris, T. M. (1989) Preparation and characterization of an aflatoxin B1 adduct with the oligodeoxynucleotide d(ATCGAT). J. Am. Chem. Soc. 111, 7232-7239. Mao, H., Deng, Z., Wang, F., Harris, T. M., and Stone, M. P. (1998) An intercalated and thermally stable FAPY adduct of aflatoxin B1 in a DNA duplex: Structural refinement from 1H NMR. Biochemistry 37, 4374-4387. Jones, W. R., Johnston, D. S., and Stone, M. P. (1998) Refined structure of the doubly intercalated d(TATAFBGCATA)2 aflatoxin B1 adduct. Chem. Res. Toxicol. 11, 873-881. Gopalakrishnan, S., Harris, T. M., and Stone, M. P. (1990) Intercalation of aflatoxin B1 in two oligodeoxynucleotide adducts: Comparative 1H NMR analysis of d(ATCAFBGAT)‚ d(ATCGAT) and d(ATAFBGCAT)2. Biochemistry 29, 10438-10448. Johnston, D. S., and Stone, M. P. (1995) Refined solution structure of 8,9-dihydro-8-(N7-guanyl)-9-hydroxy-aflatoxin B1 opposite CpA in the complementary strand of an oligodeoxynucleotide duplex as determined by 1H NMR. Biochemistry 34, 14037-14050. Bailey, E. A., Iyer, R. S., Harris, T. M., and Essigmann, J. M. (1996) A viral genome containing an unstable aflatoxin B1-N7guanine DNA adduct situated at a unique site. Nucleic Acids Res. 24, 2821-2828. Bailey, E. A., Iyer, R. S., Stone, M. P., Harris, T. M., and Essigmann, J. M. (1996) Mutational properties of the primary aflatoxin B1-DNA adduct. Proc. Natl. Acad. Sci. U.S.A. 93, 15351539. Gopalakrishnan, S., Liu, X., and Patel, D. J. (1992) Solution structure of the covalent sterigmatocystin-DNA adduct. Biochemistry 31, 10790-10801. Gopalakrishnan, S., and Patel, D. J. (1993) Formation and structural features of a sterigmatocystin-formamidopyrimidine adduct at the DNA duplex level. J. Am. Chem. Soc. 115, 93219322. Greenblatt, M. S., Bennett, W. P., Hollstein, M., and Harris, C. C. (1994) Mutations in the p53 tumor suppressor gene: Clues to cancer etiology and molecular pathogenesis. Cancer Res 54, 48554878. Shen, H. M., and Ong, C. N. (1996) Mutations of the p53 tumor suppressor gene and ras oncogenes in aflatoxin hepatocarcinogenesis. Mutat. Res. 366, 23-44. Soini, Y., Chia, S. C., Bennett, W. P., Groopman, J. D., Wang, J. S., DeBenedetti, V. M., Cawley, H., Welsh, J. A., Hansen, C., Bergasa, N. V., Jones, E. A., DiBisceglie, A. M., Trivers, G. E., Sandoval, C. A., Calderon, I. E., Munoz Espinosa, L. E., and Harris, C. C. (1996) An aflatoxin-associated mutational hotspot at codon 249 in the p53 tumor suppressor gene occurs in hepatocellular carcinomas from Mexico. Carcinogenesis 17, 10071012. Lunn, R. M., Zhang, Y. J., Wang, L. Y., Chen, C. J., Lee, P. H., Lee, C. S., Tsai, W. Y., and Santella, R. M. (1997) P53 mutations,

Chem. Res. Toxicol., Vol. 12, No. 8, 1999 713

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

chronic hepatitis B virus infection, and aflatoxin exposure in hepatocellular carcinoma in Taiwan. Cancer Res. 57, 3471-3477. Mace, K., Aguilar, F., Wang, J. S., Vautravers, P., Gomez-Lechon, M., Gonzalez, F. J., Groopman, J., Harris, C. C., and Pfeifer, A. M. (1997) Aflatoxin B1-induced DNA adduct formation and p53 mutations in cyp450-expressing human liver cell lines. Carcinogenesis 18, 1291-1297. Santella, R. M., Zhang, Y. J., Chen, C. J., Hsieh, L. L., Lee, C. S., Haghighi, B., Yang, G. Y., Wang, L. W., and Feitelson, M. (1993) Immunohistochemical detection of aflatoxin B1-DNA adducts and hepatitis B virus antigens in hepatocellular carcinoma and nontumorous liver tissue. Environ. Health Perspect. 99, 199-202. Chen, C. J., Yu, M. W., Liaw, Y. F., Wang, L. W., Chiamprasert, S., Matin, F., Hirvonen, A., Bell, D. A., and Santella, R. M. (1996) Chronic hepatitis B carriers with null genotypes of glutathione S-transferase m1 and t1 polymorphisms who are exposed to aflatoxin are at increased risk of hepatocellular carcinoma. Am. J. Hum. Genet. 59, 128-134. Yu, M. W., Lien, J. P., Chiu, Y. H., Santella, R. M., Liaw, Y. F., and Chen, C. J. (1997) Effect of aflatoxin metabolism and DNA adduct formation on hepatocellular carcinoma among chronic hepatitis B carriers in Taiwan. J. Hepatol. 27, 320-330. Kobertz, W. R., and Essigmann, J. M. (1996) Total synthesis of a cis-syn 2-carbomethoxypsoralen furan-side thymidine monadduct. J. Am. Chem. Soc. 118, 7101-7107. Kobertz, W. R., and Essigmann, J. M. (1997) Solid-phase synthesis of oligonucleotides containing a site-specific psoralen derivative. J. Am. Chem. Soc. 119, 5960-5961. Kobertz, W. R., Wang, D., Wogan, G. N., and Essigmann, J. M. (1997) An intercalation inhibitor altering the target selectivity of DNA damaging agents: synthesis of site-specific aflatoxin B1 adducts in a p53 mutational hotspot. Proc. Natl. Acad. Sci. U.S.A. 94, 9579-9584. Stone, M. P., Gopalakrishnan, S., Harris, T. M., and Graves, D. E. (1988) Carcinogen-nucleic acid interactions: Equilibrium binding studies of aflatoxins B1 and B2 with DNA and the oligodeoxynucleotide d(ATGCAT)2. J. Biomol. Struct. Dyn. 5, 1025-1041. Gopalakrishnan, S., Byrd, S., Stone, M. P., and Harris, T. M. (1989) Carcinogen-nucleic acid interactions: equilibrium binding studies of aflatoxin B1 with the oligodeoxynucleotide d(ATGCAT)2 and with plasmid pBR322 support intercalative association with the B-DNA helix. Biochemistry 28, 726-734. Raney, K. D., Gopalakrishnan, S., Byrd, S., Stone, M. P., and Harris, T. M. (1990) Alteration of the aflatoxin cyclopentenone ring to a δ-lactone reduces intercalation with DNA and decreases formation of guanine N7 adducts by aflatoxin epoxides. Chem. Res. Toxicol. 3, 254-261. Raney, V. M., Harris, T. M., and Stone, M. P. (1993) DNA conformation mediates aflatoxin B1-DNA binding and the formation of guanine N7 adducts by aflatoxin B1-8,9-exo-epoxide. Chem. Res. Toxicol. 6, 64-68. Iyer, R. S., Coles, B. F., Raney, K. D., Thier, R., Guengerich, F. P., and Harris, T. M. (1994) DNA adduction by the potent carcinogen aflatoxin B1: Mechanistic studies. J. Am. Chem. Soc. 116, 1603-1609. Strobel, S. A., Moser, H. E., and Dervan, P. B. (1988) Doublestrand cleavage of genomic DNA at a single site by triple-helix formation. J. Am. Chem. Soc. 110, 7927-7929. Moser, H. E., and Dervan, P. B. (1987) Sequence-specific cleavage of double helical DNA by triple helix formation. Science 238, 645650. Strobel, S. A., and Dervan, P. B. (1991) Single-site enzymatic cleavage of yeast genomic DNA mediated by triple helix formation. Nature 350, 172-174. Strobel, S. A., Doucette-Stamm, L. A., Riba, L., Housman, D. E., and Dervan, P. B. (1997) Site-specific cleavage of human chromosome 4 mediated by triple helix formation. Science 254, 16391642. Jones, W. R., and Stone, M. P. (1998) Site-specific targeting of aflatoxin adduction directed by triple helix formation in the major groove of oligodeoxyribonucleotides. Nucleic Acids Res. 26, 10701075. Koshlap, K. M., Gillespie, P., Dervan, P. B., and Feigon, J. (1993) Nonnatural deoxyribonucleoside D3 incorporated in an intramolecular DNA triplex binds sequence-specifically by intercalation. J. Am. Chem. Soc. 115, 7908-7909. Radhakrishnan, I., Patel, D. J., Priestly, E. S., Nash, H. M., and Dervan, P. B. (1993) NMR structural studies on a nonatural deoxyribonucleoside which mediates recognition of GC base pairs in pyrimidine-purine-pyrimidine DNA triplexes. Biochemistry 32, 11228-11234. Wang, E., Koshlap, K. M., Gillespie, P., Dervan, P. B., and Feigon, J. (1996) Solution structure of a pyrimidine-purine-pyrimidine

714 Chem. Res. Toxicol., Vol. 12, No. 8, 1999

(49) (50) (51)

(52) (53)

(54)

triplex containing the sequence-specific intercalating non-natural base D3. J. Mol. Biol. 257, 1052-1069. Maxam, A. M., and Gilbert, W. (1980) Sequencing end-labeled DNA with base-specific chemical cleavage. Methods Enzymol. 65, 499-560. Muro, A. F., Pesce, C. G., and Kornblihtt, A. R. (1993) DNA sequencing by the chemical method. Trends Genet. 9, 337-338. Stone, M. P., Gopalakrishnan, S., Raney, K. D., Raney, V. M., Byrd, S., and Harris, T. M. (1990) Aflatoxin-DNA binding and the characterization of aflatoxin B1-oligodeoxynucleotide adducts by 1H NMR spectroscopy. In Molecular Basis of Specificity in Nucleic Acid-Drug Interactions (Pullman, B., and Jortner, J., Eds.) pp 451-480, Kluwer Academic Publishers, Dordrecht, The Netherlands. Raney, K. D., Coles, B., Guengerich, F. P., and Harris, T. M. (1992) The endo-8,9-epoxide of aflatoxin B1: A new metabolite. Chem. Res. Toxicol. 5, 333-335. Johnson, W. W., Harris, T. M., and Guengerich, F. P. (1996) Kinetics and Mechanism of hydrolysis of aflatoxin B1 exo-8,9epoxide and rearrangement of the dihydrodiol. J. Am. Chem. Soc. 118, 8213-8220. Muench, K. F., Misra, R. P., and Humayun, M. Z. (1983) Sequence specificity in aflatoxin B1-DNA interactions. Proc. Natl. Acad. Sci. U.S.A. 80, 6-10.

Jones et al. (55) Benasutti, M., Ejadi, S., Whitlow, M. D., and Loechler, E. L. (1988) Mapping the binding site of aflatoxin B1 in DNA: Systematic analysis of the reactivity of aflatoxin B1 with guanines in different DNA sequences. Biochemistry 27, 472-481. (56) Mancini, D., Singh, S., Ainsworth, P., and Rodenhiser, D. (1997) Constitutively methylated CpG dinucleotides as mutation hot spots in the retinoblastoma gene (rb1). Am. J. Hum. Genet. 61, 80-87. (57) Denissenko, M. F., Pao, A., Tang, M., and Pfeifer, G. P. (1996) Preferential formation of benzo[a]pyrene adducts at lung cancer mutational hotspots in p53. Science 274, 430-432. (58) Jones, P. A., Rideout, W. M., Shen, J. C., Spruck, C. H., and Tsai, Y. C. (1992) Methylation, mutation and cancer. BioEssays 14, 3336. (59) Calladine, C. R. (1982) Mechanics of sequence-dependent stacking of bases in B-DNA. J. Mol. Biol. 161, 343-352. (60) Dickerson, R. E. (1983) Base sequence and helix structure variation in B and A DNA. J. Mol. Biol. 166, 419-441. (61) Basu, A. K., and Essigmann, J. M. (1988) Site-specifically modified oligodeoxynucleotides as probes for the structural and biological effects of DNA-damaging agents. Chem. Res. Toxicol. 1, 1-18.

TX990048U