Search

The potent tumorigen and mutagen (+)-7(R),8(S)-dihydroxy-9(S), 10(R)-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene ((+)-anti-BPDE) is a metabolite of benzo[a]pyrene that binds predominantly to the exocyclic amino group of guanine residues in DNA in vivo and in vitro. While the (-)-7S,8R,9R,10Senantiomer, (-)-anti-BPDE, also reacts with DNA to form similar covalent N2-deoxyguanosyl adducts, this diol epoxide is nontumorigenic and its mutagenic activities are different from those of (+)-anti-BPDE. In this work, T4 ligase-induced cyclization methods have been employed to demonstrate that the (+)-anti-[BP]-N2-dG lesions (G*) cause significantly greater amounts of bending and circularization of the one-base overhang undecamer duplex 5'-d(CACAT[G*]TACAC).d(TGTACATGTGG) than the stereoisomeric oligonucleotide duplex with G* = (-)-anti-[BP]-N2-dG. In the case of the (+)-anti-BPDE-modified oligonucleotides, the ratio of circular to linear DNA multimers reaches values of 8-9 for circle contour sizes of 99-121 base pairs, while for the (-)-anti-[BP]-N2-dG-modified DNA this ratio reaches a maximum value of only approximately 1 at 154-176 base pairs. Assuming a planar circle DNA model, the inferred bending angles for 90-92% of the observed circular ligation products range from 30 to 51 degrees per (+)-trans-anti-[BP]-N2-dG lesion and from 20 to 40 degrees per (-)-trans-anti-[BP]-N2-dG lesion. In the case of unmodified DNA, the probability of circular product formation is at least 1 order of magnitude less efficient than in the BPDE-modified sequences and about 90% of the circular products exhibit bending angles in the range of 14 -19 degrees . In the most abundant circular products observed experimentally, the bending angles are 40 degrees and 26 +/- 2 degrees per (+)-anti-[BP]- or (-)-anti-[BP]-modified 11-mer; these values correspond to a net contribution of 21-26 degrees and 5-19 degrees , respectively, to the observed overall bending per lesion. The coexistence of circular DNA molecules of different sizes and, therefore, different average bending angles per lesion, suggest that the lesions induce both torsional flexibility and flexible bends, which permit efficient cyclization, especially in the case of (+)-trans-[BP]-N2-dG adducts. The NMR characteristics of (+)-trans-[BP]-N2-dG lesion in the 11-mer duplex 5'-d(CACAT[G*]TACAC).d(GTGTACATGTG) indicate that all base pairs are intact, except at the underlined base pairs. This suggests a distortion in the normal conformation of the duplex on the 5'-side of the modified guanosine residue, which may be due to bending enhanced base pair opening and bending induced by the bulky carcinogen residue. The implications of base sequence-dependent flexibilities and conformational mobilities of anti-[BP]-N2-dG lesions on DNA replication and mutation are discussed.

The technique of high-resolution gel electrophoresis using oligodeoxyribonucleotides of known composition as model systems, offers a simple quantitative estimate of DNA damage in aqueous solution induced by ionizing radiation. The fraction of damaged DNA can be quantitatively defined in terms of the increased electrophoretic mobilities of the damaged oligonucleotides, relative to the mobility of the unirradiated and intact oligonucleotides. The usual direct strand breaks can be observed at gamma-ray dosages of 200 Gy. However, at a gamma-ray dosage of 400 Gy, only a broad background, attributed to heterogeneously and multiply damaged oligonucleotide fragments with overlapping and varying electrophoretic mobilities, can be distinguished. On the other hand, individual bands due to resolvable DNA fragments are evident even at dosages as high as 400 Gy for fission neutrons. When double-stranded oligonucleotides are exposed to gamma-ray dosages of 200 Gy, the fraction of damaged DNA approaches 30-40%. This damage can be almost completely suppressed (> 99%) if the irradiations are conducted in aqueous solutions in the presence of 0.5-1.0 mM concentrations of the thiols cysteamine or 3-(3-methylaminopropylamino)propanethiol (WR-151326). The rate constant of reaction of OH radicals with small double stranded oligonucleotides 16 base pairs long, KDNA, is found to be closer to the diffusion-controlled value (> 3 x 10(9) M-1 s-1) than the magnitudes of KDNA for the higher molecular weight, native DNA reported in the literature. These observations suggest that oligonucleotides represent more simple model systems than native DNA in solutions for studying the mechanisms of radioprotection exerted by thiols of different structures.

Human nucleotide excision repair processes carcinogen-DNA adducts at highly variable rates, even at adjacent sites along individual genes. Here, we identify conformational determinants of fast or slow repair by testing excision of N2-guanine adducts formed by benzo[a]pyrene diol epoxide (BPDE), a potent and ubiquitous mutagen that induces mainly G x C-->T x A transversions and frameshift deletions. We found that human nucleotide excision repair processes the predominant (+)-trans-BPDE-N2-dG adduct 15 times less efficiently than a standard acetylaminofluorene-C8-dG lesion in the same sequence. No difference was observed between (+)-trans- and (-)-trans-BPDE-N2-dG, but excision was enhanced about 10-fold by changing the adduct configurations to either (+)-cis- or (-)-cis-BPDE-N2-dG. Conversely, excision of (+)-cis- and (-)-cis- but not (+)-trans-BPDE-N2-dG was reduced about 10-fold when the complementary cytosine was replaced by adenine, and excision of these BPDE lesions was essentially abolished when the complementary deoxyribonucleotide was missing. Thus, a set of chemically identical BPDE adducts yielded a greater-than-100-fold range of repair rates, demonstrating that nucleotide excision repair activity is entirely dictated by local DNA conformation. In particular, this unique comparison between structurally highly defined substrates shows that fast excision of BPDE-N2-dG lesions is correlated with displacement of both the modified guanine and its partner base in the complementary strand from their normal intrahelical positions. The very slow excision of carcinogen-DNA adducts located opposite deletion sites reveals a cellular strategy that minimizes the fixation of frameshifts after mutagenic translesion synthesis.

A combined NMR-computational approach was employed to determine the solution structure of the (-)-trans-anti-[BP]dG adduct positioned opposite a -1 deletion site in the d(C1-C2-A3-T4-C5- [BP]G6-C7-T8-A9-C10-C11).d(G12-G13-T14-A15-G1 6-G17-A18-T19-G20-G21) sequence context. The (-)-trans-anti-[BP]dG moiety is derived from the binding of the (-)-anti-benzo[a]pyrene diol epoxide [(-)-anti-BPDE] to N2 of dG6 and has a 10R absolute configuration at the [BP]dG linkage site. The exchangeable and non-exchangeable protons of the benzo[a]pyrenyl moiety and the nucleic acid were assigned following analysis of two-dimensional NMR data sets in H2O and D2O solution. The solution conformation has been determined by incorporating intramolecular and intermolecular proton-proton distances defined by lower and upper bounds deduced from NOESY spectra as restraints in molecular mechanics computations in torsion angle space followed by restrained molecular dynamics calculations based on a NOE distance and intensity refinement protocol. Our structural studies establish that the aromatic BP ring system intercalates into the helix opposite the deletion site, while the modified deoxyguanosine residue is displaced into the minor groove with its face parallel to the helix axis. The intercalation site is wedge-shaped and the BP aromatic ring system stacks over intact flanking Watson-Crick dG.dC base pairs. The modified deoxyguanosine stacks over the minor groove face of the sugar ring of the 5'-flanking dC5 residue. The BP moiety is positioned with the benzylic ring oriented toward the minor groove and the distal pyrenyl aromatic ring directed toward the major groove. This conformation strikingly contrasts with the corresponding structure in the full duplex with the same 10R (-)-trans-anti-[BP]dG lesion positioned opposite a complementary dC residue [de los Santos et al. (1992) Biochemistry 31, 5245-5252); in this case the aromatic BP ring system is located in the minor groove, and there is no disruption of the [BP]dG.dC Watson-Crick base pairing alignment. The intercalation-base displacement features of the 10R (-)-trans-anti-[BP]dG adduct opposite a deletion site have features in common to those of the 10S (+)-trans-anti-[BP]dG adduct opposite a deletion site previously reported by Cosman et al. [(1994)(Biochemistry 33, 11507-11517], except that there is a nearly 180 degrees rotation of the BP residue about the axis of the helix at the base-displaced intercalation site and the modified deoxyguanosine is positioned in the opposite groove. In the 10S adduct, the benzylic ring is in the major groove and the aromatic ring systems point toward the minor groove. This work extends the theme of opposite orientations of adducts derived from chiral pairs of (+)- and (-)-anti-BPDE enantiomers; both 10S and 10R adducts can be positioned with opposite orientations either in the minor groove or at base displaced intercalation sites, depending on the presence or absence of the partner dC base in the complementary strand.

This study reports on the solution conformation of the covalent (+)-trans-anti-[BP]dG adduct (derived from the binding of the highly mutagenic and tumorigenic (+)-anti-benzo[a]pyrene diol epoxide to the N2 of deoxyguanosine) positioned opposite dC at a junctional site in the d(A1-A2-C3-[BP]G4-C5- T6-A7-C8-C9-A10-T11-C12-C13).d(G14-G15-A16-T17-+ ++G18-G19-T20-A21-G22-C23) 13/10-mer DNA sequence. The 13-mer represents the template strand containing the junction [BP]dG4 lesion while the complementary 10-mer models a primer strand which extends upto and is complementary to the modified dG4 residue. The solution conformation has been determined by initially incorporating intramolecular and intermolecular proton-proton distances defined by lower and upper bounds deduced from NOESY spectra as restraints in molecular mechanics computations in torsion angle space and subsequently through restrained molecular dynamics calculations based on a NOE distance and intensity refinement protocol. The duplex segment retains a minimally perturbed B-DNA conformation with all base pairs, including the junctional [BP]dG4.dC23 pair, in Watson-Crick hydrogen-bonded alignments. The pyrenyl ring is not stacked over the adjacent dC5.dG22 base pair but is positioned on the minor groove-side of the [BP]dG moiety and directed toward the 5'-end of the template strand. The pyrenyl ring stacks over the base of the non-adjacent dA2 residue in one direction and the sugar ring of dC23 in the other direction. The solution structure of the (+)-trans-anti-[BP]dG adduct opposite dC in the 13/10-mer in which the modified deoxyguanosine adopts an anti glycosidic torsion angle (this study) is in striking contrast to the structure of the same (+)-trans-anti-[BP]dG moiety in a 13/9-mer of the same sequence but without the dC23 residue positioned opposite the adduct site [Cosman, M., et al. (1995) Biochemistry 34, 15334-15350]. For the latter case, the aromatic portion of the BP residue stacks over the adjacent dC5.dG22 base pair, the modified deoxyguanosine adopts a syn glycosidic torsion angle and is displaced toward the major groove direction. Insights into the factors that affect the sequence and context dependent conformations of stereoisomeric [BP]dG lesions have emerged following comparison of these two structures with the minor groove conformations of the same (+)-trans-anti-[BP]dG lesion in the fully complementary 11-mer duplex [Cosman, M., et al. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 1914-1918] and in the base displaced-intercalative conformation of the 11/10-mer deletion duplex containing a -1 deletion site opposite the lesion [Cosman, M., et al. (1994) Biochemistry 33, 11507-11517]. The contributing factors where applicable include Watson-Crick base pairing at the site of the lesion, positioning of the carcinogen within the floor of the minor groove, and the tendency of the bulky hydrophobic aromatic BP residue to assume stacked or intercalative conformations.

Previous work has shown that the major adduct from the (+)-anti diol epoxide of benzo[a]pyrene (B[a]P), which forms at N2-deoxyguanosine [(+)-trans-anti-B[a]P-N2-dG], is capable of inducing either predominantely G --> T mutations ( approximately 95%) in a 5'-TGC-3 sequence context or predominantly G --> A mutations ( approximately 80%) in a 5'-CGT-3' sequence context. This is likely to be attributable to the major adduct being in a different mutagenic conformation in each case. In the next phase of this work, the questions to be addressed are what conformation is associated with what mutation and why? To help define what aspect of adduct structure is important to mutagenesis, the work herein reports on the mutations induced in a single sequence context by four stereoisomers of B[a]P-N2-dG: (+)-trans-, (+)-cis-, (-)-trans-, and (-)-cis-. The (+)-trans- and (-)-cis-adducts show a remarkably similar mutational pattern with G --> A mutations predominating ( approximately 80%). The (-)-trans- and (+)-cis-adducts also show a similar mutational pattern with a more even mixture of G --> T, G --> A, and G --> C mutations. Each of these adducts has an adduct bond and three hydroxyl groups at four consecutive saturated carbons in the B[a]P moiety of the adduct; the stereochemistry at these four positions differs in each of the adducts. The (+)-trans- and (-)-cis-adducts are a pair sharing the S configuration for the adduct bond, although they are a mirror image vis-a-vis the hydroxyl groups. The (-)-trans- and (+)-cis-adducts share the opposite adduct bond stereochemistry (R) but differ in the stereochemistry of their hydroxyl groups. Thus, there is a correlation suggesting that anti-B[a]P-N2-dG adduct mutagenesis is more dependent on the stereochemistry of the adduct bond than on the stereochemistry of the hydroxyl groups.

Spectroscopic, calorimetric, DNA cleavage, electrophoretic, and computer modeling techniques have been employed to characterize the DNA binding and topoisomerase poisoning properties of three protoberberine analogs, 8-desmethylcoralyne (DMC), 5,6-dihydro-8-desmethylcoralyne (DHDMC), and palmatine, which differ in the chemical structures of their B- and/or D-rings. DNA topoisomerase-mediated cleavage assays revealed that these compounds were unable to poison mammalian type II topoisomerase. By contrast, the three protoberberine analogs poisoned human topoisomerase I according to the following hierarchy: DHDMC > DMC > palmatine. DNA binding by all three protoberberine analogs induced negative flow linear dichroism signals as well as unwinding of the host duplex. These two observations are consistent with an intercalative mode of protoberberine binding to duplex DNA. However, a comparison of the DNA binding properties for DMC and DHDMC, which differ only by the state of saturation at the 5,6 positions of the B-ring, revealed that the protoberberine analogs do not "behave" like classic DNA intercalators. Specifically, saturation of the 5-6 double bond in the B-ring of DMC, thereby converting it to the DHDMC molecule, was associated with enhanced DNA unwinding as well as a reversal of DNA binding preference from a DNA duplex with an inaccessible or occluded minor groove {poly[d(G-C)]2} to DNA duplexes with accessible or unobstructed minor grooves {poly[d(A-T)]2 and poly[d(I-C)]2}. In addition, a comparison of the DNA binding properties for DHDMC and palmatine revealed that transferring the 11-methoxy moiety on the D-ring of DHDMC to the 9 position, thereby converting it to palmatine, was associated with a reduction in binding affinity for both duplexes with unobstructed minor grooves as well as for duplexes with occluded minor grooves. These DNA binding properties are consistent with a "mixed-mode" DNA binding model for protoberberines in which a portion of the ligand molecule intercalates into the double helix, while the nonintercalated portion of the ligand molecule protrudes into the minor groove of the host duplex, where it is thereby available for interactions with atoms lining the floor and/or walls of the minor groove. Furthermore, saturation at the 5,6 positions of the B-ring, which causes the A-ring to be tilted relative to the plane formed by rings C and D, appears to stabilize the interaction between the host duplex and the minor groove-directed portion of the protoberberine ligand. Computer modeling studies on the DHDMC-poly[d(A-T)]2 complex suggest that this interaction may involve van der Waals contacts between the ligand A-ring and backbone sugar atoms lining the minor groove of the host duplex. The hierarchy of topoisomerase I poisoning noted above suggests that this minor groove-directed interaction may play an important role in topoisomerase I poisoning by protoberberine analogs. In the aggregate, our results presented here, coupled with the recent demonstration of topoisomerase I poisoning by minor groove-binding terbenzimidazoles [Sun, Q., Gatto, B., Yu, C., Liu, A. , Liu, L. F., & LaVoie, E. J. (1995) J. Med. Chem. 38, 3638-3644], suggest that minor groove-directed ligand-DNA interactions may be of general importance in the poisoning of topoisomerase I.

A monoclonal antibody was developed against (+/-)-anti-benzo[c]phenanthrenediol epoxidemodified DNA, and sensitivity and specificity were determined by competitive enzyme-linked immuno-sorbent assay (ELISA). Antibody 10F9 has 50% inhibition in the ELISA at 50 fmol of B[c]PhDE-DNA adducts. There was weak cross-reactivity with DNA modified by (+/-)-anti-benzo[a]pyrenediol epoxide (50% inhibition at 150 pmol). Testing of oligonucleotides containing either (+)- or (-)-trans-anti-B[c]PhDE-adenine adducts indicated similar recognition of both stereoisomers. A quantitative immunoperoxidase technique with antibody 10F9 was developed using 10T1/2 cells treated with B[c]PhDE then piloted on exfoliated oral cells from five smokers and five nonsmokers. Mean staining in smokers (184 +/- 11) was 1.64-fold higher than in nonsmokers (112 +/- 9, p < 0.0001). This antibody should be useful for the detection and quantitation of B[c]PhDE-DNA adducts in cell culture and animal studies and in humans with environmental or occupational exposure to polycyclic aromatic hydrocarbons.

Mutations induced by the (+)-anti diol epoxide of benzo[a]pyrene [(+)-anti-B[a]PDE] were described previously in the supF gene of the Escherichia coli plasmid pUB3 [Rodriguez et al.(1993) Biochemistry, 32, 1759]. (+)-anti-B[a]PDE induced a complex pattern of mutations, including insertions, deletions, frameshifts, as well as base substitution mutations, which for G:C base pairs alone included a significant fraction of G:C --> T:A, A:T and C:G mutations. A variety of results suggest that most of these mutations arise from the major adduct ([+ta]-B[a]P-N2-dG), raising the question how can a single adduct induce different kinds of mutations? Our working hypothesis in this regard is that (1) an adduct can adopt multiple conformations; (2) different conformations cause different mutations; and (3) adduct conformation is controlled by various factors, such as DNA sequence context. To investigate what conformation is associated with what mutation, it is essential to find examples where [+ta]-B[a]P-N2-dG induces principally one kind of mutation as a prelude to the study in that same context of the conformation(s) potentially relevant to mutagenesis. Earlier work indicated that (+)-anti-B[a]PDE gave a preponderance of G --> A mutations in a 5'-CGT-3 sequence context, and herein it is shown that these mutations are likely to be attributable to the major adduct, since in this same sequence context [+ta]-B[a]P-N2-dG studied site specifically also induces principally G --> A mutations ( approximately 82%). Previously, [+ta]-B[a]P-N2-dG was shown to induce principally G --> T mutations (approximately 97%) in a 5'-TGC-3' sequence context. Thus, by simply altering its surrounding sequence context this adduct can give a preponderance of either G --> A or G --> T mutations. This is the most dramatic change in base substitution mutagenic specificity for an adduct described to date and illustrates that the qualitative pattern of mutagenesis by a bulky adduct can be remarkably diverse.

The aromatic diamidine, DAPI (4',6-diamidino-2-phenylindole), is used as an important biological and cytological tool since it forms highly fluorescent complexes with nucleic acid duplexes via minor groove-directed/intercalative modes of interaction. In this study, we find that DAPI binding can induce the formation of an RNA-DNA hybrid triplex that would not otherwise form. More specifically, through application of a broad range of spectroscopic, viscometric, and molecular modeling techniques, we demonstrate that DAPI intercalation induces the formation of the poly(dT).poly(rA).poly(dT) hybrid triple helix, a structure which does not form in the absence of the ligand. Using UV mixing studies, we demonstrate that, in the presence of DAPI, the poly(rA).poly(dT) duplex and the poly(dT) single strand form a 1:1 complex (a triplex) that does not form in the absence of DAPI. Through temperature-dependent absorbance measurements, we show that the poly(dT).poly(rA).poly(dT) triplex melts via two distinct transitions: initial conversion of the triplex to the duplex state, with the DAPI remaining bound, followed by denaturation of the duplex-DAPI complex to its component single strands and free DAPI. Using optical melting profiles, we show that DAPI binding enhances the thermal stability of the poly(dT).poly(rA).poly(dT) triplex, an observation consistent with the preferential binding of the ligand to the triplex versus the duplex and single-stranded states. Our differential scanning calorimetric measurements reveal melting of the DAPI-saturated poly(dT).poly(rA).poly(dT) triplex to be associated with a lower enthalpy but greater cooperativity than melting of the corresponding DAPI-saturated poly(rA).poly(dT) duplex. Our flow linear dichroism and viscometric data are consistent with an intercalative mode of binding when DAPI interacts with both the poly(dT).poly(rA).poly(dT) triplex and the poly(rA).poly(dT) duplex. Finally, computer modeling studies suggest that a combination of both stacking and electrostatic interactions between the intercalated ligand and the host nucleic acid play important roles in the DAPI-induced stabilization of the poly(dT).poly(rA).poly(dT) triplex. In the aggregate, our results demonstrate that ligand binding can be used to induce the formation of triplex structures that do not form in the absence of the ligand. This triplex-inducing capacity has potentially important implications in the design of novel antisense, antigene, antiviral, and diagnostic strategies.

In the supF gene, most (+)-anti-benzo[a]pyrene diol epoxide ((+)-anti-B[a]PDE) mutagenesis hot spots in Escherichia coli are in 5'-GG sequences [Rodriguez and Loechler (1993) Carcinogenesis 14, 373-383]. A major hot spot was detected at G1 in the sequence 5'-GCG1G2-CCAAAG, whereas G2 yielded very few mutants. In order to investigate the details of such sequence context effects of (+)-anti-B[a]PDE mutagenesis, we have constructed 25-mer oligonucleotides and single-stranded M13 genomes containing the above decamer sequence, in which the trans-N2-dG adduct induced by (+)-anti-B[a]PDE [(+)-trans-anti-B[a]P-N2-dG] at G1 or G2 was introduced. In vitro DNA synthesis on the adducted 25-mers was strongly blocked at each site, although the 3'-->5' exonuclease-deficient Klenow fragment could incorporate a nucleotide opposite the adduct in the presence of Mn2+. For both sites purine nucleotides were preferred. The ratio Vmax/K(m) indicated that the efficiency of incorporation of dGTP opposite these sites was very similar, but dATP incorporation opposite the adduct at G1 was five-fold more efficient than that at G2. For each site, further extension beyond the adducted nucleotide was investigated by annealing four different primers, in which only the nucleotide opposite the adducted deoxyguanosine was altered. Significant extension was only observed when deoxyadenosine was located opposite adducted G1. When the M13 genomes containing the (+)-trans-anti-B[a]P-N2-dG were replicated in E. coli, survival of each adducted genome was less than 1% as compared to the unadducted genome. Upon induction of SOS, viability increased 2-6-fold. DNA sequencing showed no base substitutions in the progeny from SOS-uninduced cells, although small deletions in a quasipalindromic sequence occurred with the adduct being located at either site. However, following SOS induction, up to 40% targeted base substitutions were detected when the adduct was located at G1, while approximately 12% of the progeny were mutants with the adduct at G2. Most base substitutions were targeted G-->T transversions. We conclude that (+)-trans-anti-B[a]P-N2-dG is a highly mutagenic and replication blocking lesion. In addition, the biological consequence of this adduct depends on whether it is located at G1 or G2, suggesting that sequence context plays a major role in the mutagenic processing of this adduct.

The Escherichia coli UvrB and UvrC proteins play key roles in DNA damage processing and incisions during nucleotide excision repair. To study the DNA structural requirements and protein-DNA intermediates formed during these processes, benzo[a]pyrene diol epoxide-damaged and structure-specific 50-base pair substrates were constructed. DNA fragments containing a preexisting 3' incision were rapidly and efficiently incised 5' to the adduct. Gel mobility shift assays indicated that this substrate supported UvrA dissociation from the UvrB-DNA complex, which led to efficient incision. Experiments with a DNA fragment containing an internal noncomplementary 11-base region surrounding the benzo[a]pyrene diol epoxide adduct indicated that UvrABC nuclease does not require fully duplexed DNA for binding and incision. In the absence of UvrA, UvrB (UvrC) bound to an 11-base noncomplementary region containing a 3' nick (Y substrate), forming a stable protein-DNA complex (Kd approximately 5-10 nM). Formation of this complex was absolutely dependent upon UvrC. Addition to this complex of ATP, but not adenosine 5'-(beta,gamma-iminotriphosphate) or adenosine 5'-(beta, gamma-methylene)triphosphate, caused incision three or four nucleotides 5' to the double strand-single strand junction. The ATPase activity of native UvrB is activated upon interaction with UvrC and enhanced further by the addition of Y substrate. Incision of this Y structure occurs even without DNA damage. Thus the UvrBC complex is a structure-specific, ATP-dependent endonuclease.

DNA replication and transcription are affected adversely by the presence of bulky adducts that are generated by the covalent binding of a variety of metabolically activated environmental pollutants to cellular DNA. When these lesions are not cleared by cellular repair enzymes prior to replication, mutations and ultimately tumor initiation can occur. Transcription and DNA repair appear to be intimately connected, since certain adducts are more efficiently removed from the transcribed strands of active loci than from non-transcribed strands and other quiescent domains in the genome. The mechanism by which RNA polymerases deal with bulky adducts during DNA transcription is therefore of great interest. The availability of site-specifically modified and stereochemically defined oligodeoxyribonucleotides derived from the covalent reaction of 7r, 8t-dihydroxy-9, 10t-epoxy- 7,8,9,10-tetrahydrobenzo[a]pyrene (anti-BPDE) with guanine residues prompted us to study the efficiencies of transcription past these lesions using bacteriophage T7 RNA polymerase. We show here that T7 RNA polymerase can bypass such lesions in a DNA template, providing that a cytosine residue is incorporated opposite anti-BPDE-modified guanine. However, when an incorrect base (most frequently a purine) is inserted opposite the modified site, the RNA polymerase stalls, and the complex dissociates, resulting in a truncated transcript. The ability of the T7 RNA polymerase to discriminate between a correct and an incorrect inserted base and, accordingly, to continue or terminate transcription, might constitute an important mechanism that ensures the fidelity of transcription past a modified base present on the transcribed strand of the DNA template.

The direct synthesis and characterization of site-specific adducts derived from the binding of (+)-1R,2S-dihydroxy-3S,4R-epoxide-1,2,3,4-tetrahydro-5-methylchrysene and the (-)-1S,2R,3R,4S-enantiomer [(+)- and (-)-5-MeCDE, respectively], to the N2-guanine residues in the oligonucleotide d(CCATCGCTACC) are described. The spectroscopic characteristics of the 5-MeCDE-modified oligonucleotides are discussed, and it is shown that their CD characteristics can be used to distinguish between the trans-addition products of the binding of the (+)- and (-)-enantiomers of 5-MeCDE (C4 position). The 11-mer duplexes with the normal complementary strands are destabilized by the site-specific, covalently bound 5-MeCDE residues: the melting points, Tm, are 5-10 degrees lower than in the case of the unmodified duplex. Stereoselective exonuclease enzyme digestion patterns of the single-stranded (+)- and (-)-trans-5-MeCDE-modified oligonucleotides (Mao et al, 1993, Biochemistry, 32, 11785-11793) were used to probe the orientations of the covalently bound 5-MeCDE residues relative to the modified guanine and the 5'-3' strand polarity; the aromatic residues are positioned either on the 5'-side [(+)-5-MeCDE], or the 3'-side [(-)-5-MeCDE adduct] of the modified guanine residues. The electrophoretic mobilities of the (+)-5-MeCDE-modified 11-mer duplexes in native polyacrylamide gels are slower than those of unmodified and modified duplexes containing the stereoisomeric (-)-5-MeCDE-N2-dG lesions. This indicates that the lesions derived from the tumorigenic (+)-5-MeCDE induce greater degrees of bending or local flexibility than the non-tumorigenic (-)-5- MeCDE enantiomer. These differences in the orientational and structural characteristics are similar to those observed with analogous DNA adducts derived from the tumorigenic (+)-7R,8S-dihydroxy-9S,10R-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene and the non-tumorigenic 7S,8R,9R,10S-enantiomer, respectively. The adducts derived from BPDE and 5-MeCDE enantiomers thus display similar characteristics that depend primarily on the PAH diol epoxide enantiomer stereochemistry. This direct synthesis approach can be used to generate milligram quantities of site-specific 5-MeCDE-modified oligonucleotides that are suitable for NMR studies (Cosman, et al., 1995, Biochemistry, 34, 6247-6260).

This paper reports on the combined NMR-molecular mechanics computational studies of the solution structure of the (-)-cis-anti-[BP]dG adduct positioned opposite dC in the sequence context d(C1- C2-A3-T4-C5-[BP]G6-C7-T8-A9-C10-C11).d(G12-G13-T14- A15-G16-C17-G18-A19-T20- G21-G22) duplex [designated (-)-cis-anti-[BP]dG.dC 11-mer duplex]. This adduct is derived from cis addition at C10 of (-)-anti-7(S),8(R)-dihydroxy-9(R),10(S)-epoxy-7,8,9,10- tetrahydrobenzo[a]pyrene [(-)-anti-BPDE] to the N2 position of dG6 in this duplex sequence. The exchangeable and nonexchangeable protons of the benzo[a]pyrenyl moiety and nucleic acid of the major conformation were assigned following analysis of two-dimensional NMR data sets in H2O and D2O solution. There was a general broadening of proton resonances for a three-nucleotide segment centered about the lesion site which resulted in a tentative assignment for the sugar protons of the C7 residue in the spectrum of the adduct duplex. The solution conformation of the major conformation of the (-)-cis-anti-[BP]dG.dC 11-mer duplex has been determined by incorporating DNA-DNA and intermolecular BP-DNA proton-proton distances defined by lower and upper bounds deduced from NOESY data sets as restraints in molecular mechanics computations in torsion angle space. The results establish that the covalently attached benzo[a]pyrenyl ring intercalates between intact Watson-Crick dC5.dG18 and dC7.dG16 base pairs. The modified deoxyguanosine [BP]-dG6 and its partner cytosine dC17 are looped out of the helix into the major groove. The purine ring of the [BP]dG6 residue is directed toward the 5'-end of the modified strand and stacks over the major groove edge of its 5'-side neighbor dC5 residue. The solution structure of the (-)-cis-anti-[BP]dG.dC 11-mer duplex is compared with those of the stereoisomeric (+)-trans-anti-[BP]dG [Cosman, M., et al. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 1914-1918], (-)-trans-anti-[BP]dG [de los Santos, C., et al. (1992) Biochemistry 31, 5245-5252], and (+)-cis-anti-[BP]dG [Cosman, M., et al. (1993a) Biochemistry 32, 4146-4155] adducts positioned opposite dC in the same duplex sequence context. A key finding is that the long axes of the intercalated benzo[a]pyrenyl rings in the solution structures of the (+)- and (-)- cis-anti-[BP]dG.dC 11-mer duplexes are oriented in opposite directions with the benzylic ring directed toward the minor groove in the (+)-cis isomer and toward the major groove in the (-)-cis isomer. In addition, a comparison is also made with the solution structure of the (+)-trans-anti-[BP]dG adduct opposite a deletion site [Cosman, M., et al. (1994a) Biochemistry 33, 11507-11517] since this adduct duplex displays several conformational features in common with the structure of the (-)-cis-anti-[BP]dG.dC 11-mer duplex. The structures of both duplex adducts exhibit intercalation of the covalently attached ligand into the helix and displacement of the modified deoxyguanosine into the major groove. Studies of the biological activities of stereochemically defined BP-DNA adducts and the comparison of the solution structure of the (-)-cis-anti-[BP]dG.dC 11-mer duplex with its stereoisomeric counterparts should lead to new insights into the relationships between defined helical distortions and mutagenic specificity and activity.

Previous studies indicated that DNA adducts formed by the carcinogenic diol epoxide 7r,8t-dihydroxy-9t,10t-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) can increase the affinity of the transcription factor Sp1 for DNA sequences that are not normally specific binding sites. Whether adducts that form in the normal binding site, the GC box sequence, increase the affinity of Sp1 for the modified GC-box was not determined. Starting with a 23-nt sequence that contains two natural GC box sequences, site-specifically modified oligonucleotides were prepared with a single(+)-BPDE-deoxyguanosine adduct at one of three positions: the center of each GC-box or in between the two boxes. Four modified oligonucleotides were studied, two derived from cis addition of BPDE to the exocyclic amino group and two from trans addition. For three of these site-specifically modified oligonucleotides, there was a diminution in Sp1 affinity, whereas Sp1 binding to the fourth modified oligonucleotide was abolished. Furthermore, random modification of the oligonucleotide to a level of about 1 BPDE adduct per fragment slightly decreased the affinity for Sp1, and no evidence was found for a subpopulation of molecules with high affinity. These findings suggest that BPDE modification of the GC box does not lead to an increased affinity for Sp1. This is consistent with a model in which a BPDE-induced bend in the DNA mimics the conformation of the normal GC box:Sp1 complex, leading to high-affinity binding of Sp1 to non-Gc box sites.

We have used a broad range of spectroscopic and viscometric techniques to demonstrate that the complexation of a cytotoxic, topoisomerase I-poisoning terbenzimidazole (5PTB) with the poly(dA).poly(dT) duplex exhibits properties characteristic of both intercalation and minor groove binding. Our results reveal the following features: (i) Optical melting profiles reveal that 5PTB binding enhances the thermal stability of the poly(dA).poly(dT) duplex; (ii) Fluorescence-detected 5PTB binding to the poly(dA).poly(dT) duplex reveals four apparent "site sizes," ranging from 1 to 13 base pairs (bp) per bound drug; (iii) Flow linear dichroism data suggest conformational heterogeneity among the poly(dA).poly(dT)-bound 5PTB molecules, with substantial contributions from drug molecules bound in the minor groove; (iv) Fluorescence resonance energy transfer data reveal properties characteristic of a significant contribution from an intercalative mode of binding; (v) Viscometric, fluorescence quenching, and netropsin competition data are consistent with 5PTB binding to poly(dA).poly(dT) by "mixed" modes, which are operationally defined as single or multiple binding populations that individually and/or collectively express both intercalative and minor groove binding properties. We comment on a potential correlation between drugs that exhibit such "mixed" mode binding motifs and those that express antineoplastic activity through inhibition of topoisomerase I.

The site specifically modified oligonucleotides 5'-d(TCCTCCTG1G2CCTCTC) (I) and 5'-d(CTATG1G2G3TATC) (II) were synthesized with single modified guanine residues at positions G1, G2, or G3, derived from the covalent binding reaction of 7R,8S-dihydroxy-9S,10R-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene ((+)-anti-BPDE) with the exocyclic amino groups of the guanine residues. In denaturing 20% polyacrylamide gels, the electrophoretic mobilities of the (+)-anti-BPDE-modified oligonucleotides I and II are slower than the mobilities of the respective unmodified oligonucleotides and independent of the positions of the BPDE-modified guanines. However, in the double-stranded forms in native 8% polyacrylamide gels, the electrophoretic mobilities of the duplexes with lesions at G2 or G3 are remarkably slower (reductions in mobilities up to approximately 40%) than to duplexes with lesions at G1 and are attributed to physical bends or flexible hinge joints at the sites of the BPDE lesions. These sequence-dependent mobility effects occur whenever the BPDE-modified guanine residues with (+)-trans-stereochemistry are flanked by unmodified G's on the 5'-side. These retarded electrophoretic mobilities are attributed to bending induced by steric hindrance effects involving the bulky 5'-flanking guanines and the pyrenyl residues that are known to point into the 5'-direction relative to the modified G [Cosman, M., et al. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 1914-1918]. These anomalous electrophoretic mobility effects are not observed in the case of (-)-anti-BPDE-modified sequences I with trans-(-)-anti-BPDE-N2-dG adduct stereochemistry.

The structural features of a chemically modified DNA template strand may promote error-prone DNA synthesis during replication. The resulting higher incidence of mutations, in turn, can eventually lead to tumor initiation. Structural insights into this process can be monitored by studying chemically modified base adducts of defined stereochemistry positioned site-specifically at a single strand--duplex template--primer junction. We have used a NMR-molecular mechanics approach to obtain the solution conformations of the covalent adducts derived from trans additions at the [BP]C10 position of the highly tumorigenic (+)-anti-benzo[a]pyrene diol epoxide [(+)-anti-BPDE] and nontumorigenic (-)-anti-benzo-[a]pyrene diol epoxide [(-)-anti-BPDE] to the N2 position of guanine [(+) and (-)-trans-anti-[BP]dG, respectively] in the d(A1-A2-C3-[BP]G4-C5-T6-A7-C8-C9-A10-T11-C12-C13).d (G14-G15-A16-T17-G18-G19-T20-A 21-G22) 13/9-mer DNA sequence. The modified 13-mer strand constitutes the template strand, while the complementary 9-mer strand constitutes a primer which has been synthesized from the 3'-end of the template toward the 5'-end up to the base preceding, but not including, the modified guanine. The modified guanine (denoted by [BP]dG4) is positioned at the junction site between the single-stranded and duplex segments. Structural features of the (+)-trans-anti-[BP]dG 13/9-mer have been determined by incorporating proton--proton distances defined by lower and upper bounds deduced from NOESY spectra as restraints in molecular mechanics computations in torsion angle space. The 3'-side duplex segment retains a minimally perturbed B-DNA conformation with all nine base pairs in Watson--Crick hydrogen-bonded alignments. Conformational heterogeneity is detected at the single-stranded d(A1-A2-C3) segment located 5' to the modified (+)-trans-anti-[BP]dG lesion which contrasts with an unperturbed alignment of these same residues in the unmodified control 13/9-mer. The modified guanine adopts a syn glycosidic torsion angle, is displaced into the major groove, and no longer stacks over the adjacent dC5.dG22 base pair. Such a base displacement is accompanied by stacking of one face of the pyrenyl ring with the dC5.dG22 base pair located on the duplex segment proximate to the modified guanine, while the other face of BP is exposed to solvent.(ABSTRACT TRUNCATED AT 400 WORDS)

Using polyacrylamide gel electrophoresis (PAGE) and low-temperature, laser-induced fluorescence line narrowing (FLN) and non-line narrowing (NLN) spectroscopic methods, the conformational characteristics of stereochemically defined and site-specific adducts derived from the binding of 7 beta,8 alpha-dihydroxy-9 alpha,10 alpha-epoxy-7,8,9,10- tetrahydrobenzo[a]pyrene (anti-BPDE, a metabolite of the environmental carcinogen benzo[a]pyrene), to DNA were studied. The focus of these studies was on the four stereochemically distinct anti-BPDE modified duplexes 5'-d(CCATCGCTACC).(GGTAGCGATGG), where G denotes the lesion site derived from trans or cis addition of the exocyclic amino group of guanine to the C10 position of either (+) or (-)-anti-BPDE. PAGE experiments under non-denaturing conditions showed that the (+)-trans adduct causes a significantly greater retardation in the electrophoretic mobility than the other three adducts, probably the result of important adduct-induced distortions of the duplex structure. Low-temperature fluorescence studies in frozen aqueous buffer matrices showed that the (+)-trans adduct adopts primarily an external conformation with only minor interactions with the helix, but a smaller fraction (approximately 25%) appears to exists in a partially base-stacked conformation. The (-)-trans adduct exists almost exclusively (approximately 97%) in an external conformation. Both cis adducts were found to be intercalated; strong electron-phonon coupling observed in their FLN spectra provided additional evidence for significant pi-pi stacking interactions between the pyrenyl residues and the bases. FLN spectroscopy is shown to be suitable for distinguishing between trans and cis adducts, but lesions with either (+)- or (-)-trans, or (+)- or (-)-cis stereochemical characteristics showed very similar vibrational patterns.(ABSTRACT TRUNCATED AT 250 WORDS)

Treatments with psoralens and long-wavelength ultraviolet radiation (UVA, 320-400 nm; PUVA) have shown efficacy for virus sterilization of platelet concentrates (PC). Our laboratory has employed the psoralen derivative 4'-aminomethyl-4,5',8-trimethylpsoralen (AMT), and we have found that platelet integrity is best preserved when rutin, a flavonoid that quenches multiple reactive oxygen species, is present during AMT/UVA treatment of PC. In this report, we examine the effects of different UVA spectra under our standard PC treatment conditions (i.e. 50 micrograms/mL AMT, 0.35 mM rutin and 38 J/cm2 UVA). Added vesicular stomatitis virus (VSV; > or = 5.5 log10) was completely inactivated with the simultaneous maintenance of the platelet aggregation response (> 90% of control) when a UVA light source with transmission mainly between 360 and 370 nm (narrow UVA1) was used. In contrast, with a broad-band UVA (320-400 nm; broad UVA) light source, the aggregation response was greatly compromised (< 50% of control) with only a minor increase in the rate of VSV kill. With this lamp, platelet function could be improved to about 75% of the control by adding a long-pass filter, which reduced the transmission of shorter (< or = 345 nm) UVA wavelengths (340-400 nm; UVA1). At equivalent levels of virus kill, aggregation function was always best preserved when narrow UVA1 was used for PUVA treatment. Even in the absence of AMT, and with or without rutin present, narrow UVA1 irradiation was better tolerated by platelets than was broad UVA.(ABSTRACT TRUNCATED AT 250 WORDS)

Mutations induced by the (+)-anti-diol epoxide of benzo[a]pyrene [(+)-anti-B[a]PDE] were collected in the supF gene of the Escherichia coli plasmid pUB3. pUB3 was reacted with (+)-anti-B[a]-PDE and then either (1) transformed immediately into E. coli or (2) heated at 80 degrees C for 10 min and then cooled prior to transformation--the latter to probe mechanism [Rodriguez & Loechler (1993) Biochemistry 32, 1759]. Qualitatively, heating did not affect the mutagenic pattern, except at the major base substitution hotspot in supF, G115, where principally G-->T mutations were obtained prior to heating, while after heating, G-->A and G-->C mutations became statistically significantly more prevalent. Several studies have suggested that a heat-induced chemical transformation of a (+)-anti-B[a]PDE adduct at G115 (e.g., into an apurinic site) is not likely to explain the change in mutational pattern. The most likely model is that (+)-anti-B[a]P-N2-Gua is initially trapped in a metastable conformation giving principally G-->T mutations, while heating induces a change to a stable conformation(s) resulting in G-->T, A, and C mutations. This suggests that adduct conformational complexity is at the root of adduct mutational complexity. To investigate this model, a plasmid (B[a]P-G115-pRE1) with (+)-anti-B[a]P-N2-Gua in the G115 sequence context is constructed using adduct site-specific techniques.(ABSTRACT TRUNCATED AT 250 WORDS)

Oligonucleotides containing site-specifically-modified N2-guanine (+)-trans-, (-)-trans-, (+)-cis-, and (-)-cis-BPDE adducts were ligated into 50-base-pair DNA fragments. These substrates were used in reactions with the Escherichia coli UvrABC nuclease system. The interaction of the UvrA2 and UvrA2B complexes with these four stereoisomers was probed using DNase I footprinting and gel mobility shift assays. DNase I digestion of substrates containing each stereoisomer of BPDE displayed a unique pattern which was consistent with the known structure of these DNA adducts. UvrA and UvrA2B appeared to interact very similarly with all four substrates. Binding of UvrA2 to these substrates produced a 33-bp footprint, and the UvrB--DNA complex resulted in footprint of 24 bp. The UvrABC nuclease system produced bimodal incisions at the eighth phosphate 5' and the fifth, sixth, or seventh phosphate 3' to the modified guanine. The variation of the 3' incision site was linked to the stereochemistry and orientation of the BPDE adduct. For example, the 3' incision of the 50-bp duplex containing (-)-trans-BPDE-N2-guanine was inhibited at the fifth phosphate. UvrABC nuclease incision kinetics revealed a hierarchy of specificity. The intercalative cis isomers were incised more efficiently than the corresponding trans isomers which lie in the minor groove. The (+) enantiomers were incised more efficiently than the (-) form for both cis and trans isomers. These observations reveal that UvrABC nuclease recognition and incision are directly influenced by the conformation of the DNA adduct.

Antisera developed against benzo[a]pyrene diolepoxide (BPDE)-DNA adducts are sensitive tools for detection of DNA adducts in human samples. All antisera currently used for biomonitoring studies were produced against DNA or guanosine modified with racemic anti-BPDE. Using a non-competitive enzyme-linked immunosorbent assay (ELISA), Venkatachalam and Wani (Carcinogenesis, 15, 565-572, 1994) recently tested polyclonal and monoclonal (5D2) antisera for cross-reactivity against oligonucleotides containing (+)-and (-)-trans-anti-BPDE-N2-guanine or N6-adenine adducts and showed different stereospecificity for the two antisera. Because of the importance of antiserum specificity in human biomonitoring studies, we have tested several monoclonal (Mab 5D11 and 5D2) and polyclonal (Pab #29) antisera developed against racemic anti-BPDE-DNA adducts, and Mab 8E11 developed against anti-BPDE-guanosine adducts. Stereoisomeric anti-BPDE-modified oligonucleotide adducts in the sequence 5'-d(CC-AT-CG*CTACC)-3' where G* = anti-BPDE-N2-dG with (+)-trans, (-)-trans, (+)-cis and (-)-cis adduct stereochemistry at the C10 position of anti-BPDE were tested by competitive ELISA. Two structurally related 5-methylchrysene diolepoxide adducts with G* = (+)- and (-)-trans-anti-5-MeCDE-N2-dG in the same oligonucleotide were also tested. While Mab5D2 had the highest affinity for the (-)-trans-anti-BPDE-modified oligomer, Mab 5D11 and 8E11 and Pab #29 recognized the (+)-trans-anti-BPDE-modified oligomer better than the (-)-trans-anti-BPDE modified oligomer. Mab 5D11 and Pab #29 recognized racemic anti-BPDE-modified DNA adducts better than trans-anti-BPDE-modified oligonucleotides; however, Mab 8E11 showed similar sensitivity to racemic anti-BPDE-DNA adducts and (+)- and (-)-trans-anti-BPDE-modified oligomers. All antisera exhibited lower reactivities with both 5-MeCDE modified oligomers. Because of their sensitive detection of (+)-trans-anti-BPDE-dG adducts, the primary adduct produced in vivo, Mab 8E11 and 5D11 and Pab #29 are appropriate for measurement of most adducts formed in humans.

The apparent persistence length of enzymatically linearized pIBI30 plasmid DNA molecules approximately 2300 bp long, as measured by a hydrodynamic linear flow dichroism method, is markedly decreased after covalent binding of the highly tumorigenic benzo[a]pyrene metabolite 7R,8S-dihydroxy-9S,10R-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene [(+)-anti-BPDE]. In striking contrast, the binding of the non-tumorigenic, mirror-image 7S,8R,9R,10S enantiomer [(-)-anti-BPDE] to DNA has no measurable effect on its alignment in hydrodynamic flow gradients (< or = 2.2% of the DNA bases modified). In order to relate this effect to BPDE-nucleotide lesions of defined stereochemistry, the bending induced by site-specifically placed and stereochemically defined (+)- and (-)-anti-BPDE-N2-dG lesions in an 11mer deoxyoligonucleotide duplex was studied by ligation and gel electrophoresis methods. Out of the four stereochemically isomeric anti-BPDE-N2-deoxyguanosyl (dG) adducts with either (+)-trans, (-)-trans, (+)-cis, and (-)-cis adduct stereochemistry, only the (+)-trans adduct gives rise to prominent bends or flexible hinge joints in the modified oligonucleotide duplexes. Since both anti-BPDE enantiomers are known to bind preferentially to dG (> or = 85%), these observations can account for the differences in persistence lengths of DNA modified with either (+)-anti-BPDE or the chiral (-)-anti-BPDE isomer.

This paper reports on NMR-molecular mechanics structural studies of the (-)-trans-anti-[MC]dG adduct positioned opposite dC in the sequence context of the d(C1-C2-A3-T4-C5-[MC]G6-C7-T8-A9-C10-C11).d(G12-G13-T14++ +-A15-G16-C17-G18- A19-T20-G21-G22) duplex [designated (-)-trans-anti-[MC]dG.dC 11-mer duplex]. This adduct is derived from the trans addition at C4 of (-)-anti-1(S),2(R)-dihydroxy-3(R),4(S)-epoxy-1,2,3,4-tetrahydro-5-met hylchrysen e [(-)-anti-5-MeCDE] to the N2 position of dG6 in this duplex sequence. The 5-methyl group is located adjacent to the MC(C4) binding site, with these groups juxtaposed in a sterically crowded bay region in the adduct duplex. The 5-methylchrysenyl and the nucleic acid exchangeable and nonexchangeable protons were assigned following analysis of two-dimensional NMR data sets in H2O and D2O buffer solution. The solution structure of the (-)-trans-anti-[MC]dG.dC 11-mer duplex has been determined by incorporating DNA-DNA and carcinogen-DNA proton-proton distances defined by lower and upper bounds deduced from NOESY data sets as restraints in molecular mechanics computations in torsion angle space. The results establish that the [MC]dG6.dC17 base pair and flanking dC5.dG18 and dC7.dG16 base pairs retain Watson-Crick alignments upon adduct formation. The aromatic chrysenyl ring is positioned in the minor groove of a right-handed B-DNA helix and stacks predominantly over the sugar of the dC17 residue across from it on the unmodified complementary strand. The chrysenyl ring points toward the 3'-end of the modified strand with its 5-methyl group inserting between the modified [MC]dG6.dC17 and dC7.dG16 base pairs. The adduct duplex bends by approximately 47 degrees as a result of the wedged insertion of the 5-methyl group from the minor groove face of the duplex. The solution structure of the (-)-trans-anti-[MC] dG.dC 11-mer duplex is compared with that of the corresponding (-)-trans-anti-[BP]dG.dC 11-mer [De los Santos et al. (1992) Biochemistry 31, 5245-5252] in which the [BP]dG adduct is derived from the binding of (-)-anti-BPDE [7(S),8(R)-dihydroxy-9(R),10(S)-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene] to the N2 position in the same DNA sequence context. Although the solution structures of the (-)-trans-anti-stereoisomers of 5-methylchrysenyl-dG and benzo[a]pyrenyl-dG adducts opposite dC exhibit many features in common with each other, the [MC]dG adduct which contains a bay region methyl group bends the DNA helix to a greater extent than in the corresponding [BP]dG adduct, which lacks a bay region methyl group.(ABSTRACT TRUNCATED AT 400 WORDS)

Background: Photodynamic treatment of red cells (RBCs) with phthalocyanines and red light inactivates lipid-enveloped viruses, such as vesicular stomatitis virus (VSV) and human immunodeficiency virus. To protect RBCs from photodynamic damage, type I free radical quenchers, such as mannitol, which did not affect virus inactivation, were added., Study Design and Methods: Aluminum phthalocyanine tetrasulfonate (AIPcS4) was found to inactivate VSV at a rate one-fourth that of the silicon phthalocyanines (Pc 4 and Pc 5). However, the latter also caused more RBC damage. To protect RBCs against this photodynamic damage, Trolox, a water-soluble vitamin E analogue, was used. RBC damage was measured as potassium leakage or hemolysis during storage after treatment. In addition, reduction in negative surface charge on RBCs was measured immediately after treatment, and the effect of Trolox on VSV inactivation in RBCs was evaluated., Results: Trolox at a concentration of 5 mM was found to reduce potassium leakage during storage after Pc 4 and AIPcS4 photodynamic treatment of RBCs. Hemolysis during storage of RBC concentrates treated with Pc 4 or Pc 5 was drastically reduced by the addition of 5 mM Trolox prior to light exposure. At the same concentration, Trolox inhibited the reduction of negative surface charges on RBCs following Pc 4 and Pc 5 photodynamic treatment. Under these conditions, VSV inactivation by photodynamic treatment with all phthalocyanines was not affected by Trolox. In aqueous solution, Trolox formed a complex with AIPcS4, thus quenching the excited triplet state of AIPcS4 at a constant rate of 8.8 x 10(6) per M per second., Conclusion: These findings indicate that Trolox protects RBCs from phthalocyanine-photosensitized damage without affecting virus kill. The addition of Trolox would be beneficial for improving the quality of RBCs subjected to photodynamic treatment.

Site-specifically modified oligonucleotides were obtained in milligram quantities by reacting racemic 3t,4r-dihydroxy-1,2t-epoxy-1,2,3,4-tetrahydrobenzo[c]phenanthrene (B[c]PhDE-2, or anti-B[c]PhDE) with the single deoxyadenosine (dA) residue in the oligodeoxynucleotide d(CTCTCACTTCC). Enzyme digestion of the covalently modified oligonucleotides with the exonuclease spleen phosphodiesterase yielded covalently linked B[ca]PhDE-N6-deoxyadenosyl monophosphate (dAMP) adducts. Comparisons of the reverse phase HPLC retention times and CD spectra of these B[c]PhDE-3'-dAMP mononucleotide adducts, with those of standards derived from the reaction of the enantiomers (+)- and (-)-anti-B[c]PhDE with 3'-dAMP, show that two major oligonucleotide adducts (I and II) were obtained upon reacting racemic anti-B[c]PhDE with d(CTCTCACTTCC). In oligonucleotide adduct I, the lesion is a (+)-trans-anti-B[c]PhDE-N6-dA residue, and in oligonucleotide adduct II it is a (-)-trans-anti-B[c]PhDE-N6-dA residue. These assignments were further confirmed using a standard 32P postlabeling assay of B[c]PhDE-3'-dAMP mononucleotide adducts obtained from the digestion of oligonucleotides I and II by spleen phosphodiesterase. The melting points (Tm) of duplexes of modified oligonucleotides I and II and their natural complementary strands are not affected significantly by the presence of the covalently bound benzo[c]phenanthrenyl residues. Opposite stereoselective resistance to enzyme digestion by the exonucleases snake venom phosphodiesterase and spleen phosphodiesterase is exhibited by the stereoisomeric (+)-trans- and (-)-trans-anti-B[c]PhDE-modified oligonucleotide adducts I and II; these results are consistent with the intercalative insertion of the benzo[c]phenanthrenyl residues on the 5'-side of the modified dA residue in adduct I, and its insertion on the 3'-side of the dA residue in adduct II, as observed in the duplexes by high resolution NMR techniques [Cosman et al. (1993) Biochemistry 32, 12488-12497, and Cosman et all, Biochemistry, in press.

The specificity of the laser pulse-induced photocleavage method [Boles, T. C., and Hogan, M. E. (1986) Biochemistry 25, 3039] for detecting the sites of covalent binding of benzo[a]pyrene diol epoxide in DNA was investigated using site-specifically modified oligonucleotide duplexes d(CTCACAT[G*]TACACTCT).d(GAGAGTGTACATGTGA), where [G*] is the adducted guanine residue (+)-trans-anti-BPDE-N2-dG (anti-BPDE = 7 beta,8 alpha- dihydroxy-9 alpha,10 alpha-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene). The appropriate oligonucleotide strands were 32P-end-labeled, and the duplexes were irradiated with a pulse train of 355 nm Nd:YAG laser pulses (approximately 300 mW/cm2). The products of the photocleavage reaction were analyzed by denaturing gel electrophoresis. The major observed products included the intact oligonucleotide strand that had lost the BPDE residue, and shorter oligonucleotide fragments arising from strand scission at the BPDE-modified guanine residue and at nearby flanking bases. Photocleavage at the BPDE-modified G is dominant (approximately 50 +/- 5% of the sum of all of the shorter oligonucleotide fragments), and cleavage extends to at least 4 bases on the 5'-side, and 7 bases on the 3'-side of the BPDE-modified G residue; the probability of cleavage diminishes with increasing distance from the modified G residue. On the unmodified complementary strand, nonspecific strand cleavage is also observed, but the probability of cleavage is > or = 20 times smaller than at the BPDE-modified G residue on the modified strand. The photocleavage method thus preferentially causes strand scission at the sites of BPDE modification, but the occurrence of strand cleavage with lower probabilities at neighboring sites is also significant.

Direct synthesis and purification procedures are described for the preparation of adducts derived from the covalent binding of 7R,8S-dihydroxy-9S,10R-epoxy-7,8,9,10-tetrahydro-benzo[a]pyrene [(+)-anti-BPDE or (+)-BPDE 2] to each of the three guanine residues (trans-N2-dG lesions) in the oligodeoxyribonucleotide d(CTATG1G2G3TATC). The positions of the modified Gs are defined by Maxam-Gilbert sequencing techniques. Six different oligonucleotides with one or two precisely positioned (+)-anti-BPDE residues are identified. The absorbance, circular dichroism and fluorescence characteristics are changed upon formation of duplexes with the complementary strands d(GATACCCATAG). In the doubly-modified oligonucleotides, a broad, excimer-like long wavelength fluorescence emission band is observed with a maximum near 455 nm only if the two (+)-anti-BPDE-modified Gs are adjacent to one another. The covalently attached (+)-anti-BPDE residues decrease the thermodynamic stabilities of the duplexes; their melting points are markedly dependent on the position of the lesions, being highest with the (+)-anti-BPDE residue at G1 (Tm = 40 degrees C, only 2 degrees C lower than in the case of the unmodified oligonucleotide) and lowest when it is situated at G3 (Tm = 29 degrees C). The implications of these and other physical characteristics are discussed. The facile synthesis of these or similar site-specific and stereochemically defined (+)-trans-anti-BPDE-N2-dG lesions in runs of contiguous guanines in oligodeoxyribonucleotides of specified base sequence should be useful for the design of site-directed mutagenesis studies in vitro and in vivo.

Phthalocyanines are being studied as photosensitizers for virus sterilization of red blood cells (RBC). During optimization of the reaction conditions, we observed a marked effect of the irradiance on production of RBC damage. Using a broad-band light source (600-700 nm) between 5 and 80 mW/cm2, there was an inverse relationship between irradiance and rate of photohemolysis. This effect was observed with aluminum sulfonated phthalocyanine (AlPcSn) and cationic silicon (HOSiPc-OSi[CH3]2[CH2]3N+[CH3]3I- phthalocyanine (Pc5) photosensitizers. The same effect occurred when the reduction of RBC negative surface charges was used as an endpoint. Under the same treatment conditions, vesicular stomatitis virus inactivation rate was unaffected by changes in the irradiance. Reduction in oxygen availability for the photochemical reaction at high irradiance could explain the effect. However, theoretical estimates suggest that oxygen depletion is minimal under our conditions. In addition, because the rate of photohemolysis at 80 mW/cm2 was not increased when irradiations were carried out under an oxygen atmosphere this seems unlikely. Likewise, formation of singlet oxygen dimoles at high irradiances does not appear to be involved because the effect was unchanged when light exposure was in D2O. While there is no ready explanation for this irradiance effect, it could be used to increase the safety margin of RBC virucidal treatment by employing exposure at high irradiance, thus minimizing the damage to RBC.

This paper reports on NMR-molecular mechanics structural studies of the (-)- trans-anti-benzo[c]phenanthrene-dA adduct positioned opposite dT in the sequence context of the d(C1-T2-C3-T4-C5-[BPh]A6-C7-T8-T9-C10-C11).d(G12- G13-A14-A15-G16-T17-G18-A19-G20-A21- G22) duplex (designated as the (-)-trans-anti-[BPh]dA.dT 11-mer duplex). This adduct is derived from the covalent binding of (-)-1,2-dihydroxy-3,4-epoxy-1,2,3,4-tetrahydro-benzo[c]phenanthrene [(-)-anti-BPhDE] to N6 of dA6 in this duplex sequence. The benzo[c]phenanthrenyl and nucleic acid exchangeable and nonexchangeable protons were assigned in the predominant conformation following analysis of two-dimensional NMR data sets in H2O and D2O buffer solution. The solution structure of the (-)-trans-anti-[BPh]dA.dT 11-mer duplex has been determined by incorporating intramolecular and carcinogen-DNA proton-proton distances defined by lower and upper bounds deduced from NOESY data sets as restraints in molecular mechanics computations in torsion angle space. The results show that the [BPh]dA6.dT17 base pair propeller twists and buckles slightly to permit the covalently attached benzo[c]phenanthrenyl ring to intercalate between the [BPh]dA6.dT17 and dC7.dG16 base pairs to the 3'-side of the [BPh]dA6 lesion site without disrupting the Watson-Crick hydrogen bond alignments in the modified duplex. The strain in the highly sterically hindered fjord region of the benzo[c]phenanthrenyl moiety is relieved by the propeller-like nonplanar geometry of the aromatic phenanthrenyl ring system, which stacks predominantly with the dG16 and dT17 bases on the unmodified strand. The benzylic ring adopts a distorted half-chair form, in which the H1 and H2 protons are pseudo-diequatorial and the H3 and H4 protons are pseudodiaxial. The current observation that the (-)-trans-anti-[BPh]dA positioned opposite dT intercalates to the 3'-side of the intact modified base pair contrasts with our previous demonstration that the stereoisomeric (+)-trans-anti-[BPh]dA adduct positioned opposite dT intercalates to the 5'-side of the intact modified base pair [Cosman, M., et al. (1993b) Biochemistry 32, 12488-12497]. These stereochemically induced structural differences between isomeric [BPh]dA lesions derived from the binding of chiral (+)- and (-)-anti-BPhDE enantiomers may in turn profoundly influence the interactions of the carcinogen-modified DNA with repair and replication enzymes in the cell.

Conformations of the trans adduct of (+)-anti-benzo[a]pyrene -7,8-dihydrodiol-9,10-epoxide (BPDE) to N2-guanine, the major stable DNA adduct of the environmental carcinogen benzo[a]pyrene, were studied as a function of flanking bases in single-stranded and in double-stranded oligonucleotides. Three 11mer oligonucleotides d(CTAT-G1G2G3TATC) were synthesized containing the (+)-trans-anti-BPDE adduct at one specific guanine of the GGG sequence (a known mutational hot spot). Polyacrylamide gel electrophoresis of the three single-stranded oligonucleotides showed that the adduct bound to G2 or G3 (5'-flanking base guanine) caused significantly stronger retardation than the same adduct bound to G1 (5'-flanking base thymine). The strength of the carcinogen-base interaction was reflected in the spectroscopic properties of the pyrenyl moiety. Low temperature fluorescence measurements under line-narrowing (FLN) or non-line-narrowing (NLN) conditions showed that in single-stranded form the adduct at G2 or G3 (5'-flanking base guanine) adopts a conformation with strong interaction with the bases. This was also observed for the same adduct at the sequence AGA. In contrast, the (+)-trans-anti-BPDE adduct with a 5'-flanking thymine exists in a primarily helix-external conformation. Similar differences were observed in the double-stranded oligonucleotides: the adducts at G2 and G3 were found to exist in similar conformational equilibria, again with significant carcinogen-base interactions, while the adduct at G1 showed a predominantly external conformation. The nature of the 3'-flanking base appeared to have little influence on the conformational equilibrium of the (+)-trans-anti-BPDE-guanine adduct. The results could provide insight into the mutational specificity and flanking base effects observed for (+)-anti-BPDE.

We have applied a combined NMR-molecular mechanics approach to determine the solution structure of the (+)-cis-anti-[BP]dG adduct positioned opposite a deletion site in the sequence d(C5-[BP]G6-C7).d(G16-G17) at the DNA oligomer duplex level. Our structural studies establish that the benzo[a]pyrene ring intercalates into the helix opposite the deletion site while the modified deoxyguanosine is displaced into the minor groove with its plane parallel to the helix axis. The intercalation site is wedge-shaped with the benzo[a]pyrene ring stacked over intact flanking Watson-Crick dG.dC base pairs. The modified deoxyguanosine stacks over the minor groove face of the sugar ring of the 5'-flanking dC5 residue. The structure at the lesion site is consistent with the observed intermolecular NOEs which served as input restraints to guide the molecular mechanics calculations, and, in addition, the various stacking interactions explain the observed large ring current shifts associated with adduct formation. The solution structure of the (+)-cis-anti-[BP]dG adduct positioned opposite a deletion site reported in this study is similar to the corresponding structure of the same adduct positioned opposite dC reported previously [Cosman, M., de los Santos, C., Fiala, R., Hingerty, B. E., Luna, E., Harvey, R. G., Geacintov, N. E., Broyde, S., & Patel, D. J. (1993) Biochemistry 32, 4145-4155]. This is not surprising since the dC opposite the lesion site was looped out of the helix and it can be readily replaced by a deletion site through minor changes associated with buckling of the intercalation site. By contrast, the solution structures of the (+)-trans-anti-[BP]dG and (+)-cis-anti-[BP]dG adducts positioned opposite deletion sites in the same sequence context exhibit distinct surface topologies in the grooves of the DNA helix. Thus, even though the benzo[a]pyrene intercalates into the helix opposite the deletion site in both cases, the modified deoxyguanosine is displaced into the major groove for the (+)-trans-anti-[BP]dG adduct while it is displaced into the minor groove for the (+)-cis-anti-[BP]dG stereoisomer. The orientational differences reflect the chiral characteristics of the two [BP]-dG stereoisomeric adducts with the different alignments of the bulky DNA lesion opposite the deletion site likely to influence interactions with the cellular repair machinery.

This paper reports on the solution structure of the (+)-trans-anti-[BP]dG adduct positioned opposite a deletion site in a DNA oligomer duplex which defines the alignment of this covalent benzo[a]pyrene-N2-deoxyguanosine stereosiomer relative to the deletion site. The combined NMR-molecular mechanics computation studies were undertaken on the (+)-trans-anti-[BP]dG adduct embedded in the d(C5-[BP]G6-C7).d(G16-G17) sequence context in a duplex containing 11 residues on the modified strand and 10 on the partner, with no base opposite the modification. The exchangeable and nonexchangeable protons of the benzo[a]pyrenyl moiety and the nucleic acid were assigned following analysis of two-dimensional NMR data sets in H2O and D2O solution. The solution conformation of the (+)-trans-anti-[BP]dG.del 11-mer duplex has been determined by incorporating intramolecular and intermolecular proton-proton distances defined by lower and upper bounds deduced from NOESY spectra as restraints in molecular mechanics computations in torsion angle space. The benzo[a]pyrene ring of [BP]dG6 is intercalated between intact Watson-Crick dC5.dG17 and dC7.dG16 base pairs with the deoxyguanosine base of [BP]dG6 displaced into the major groove. The intercalation site is wedge shaped, being narrower toward the dG16-dG17 step on the deletion-containing strand. The deoxyguanosine base of [BP]dG6 which is positioned in the major groove is inclined relative to the helix axis and stacks over the 5'-flanking dC5 residue in the solution structure. The intercalative-base displacement structure of the (+)-trans-anti-[BP]dG.del 11-mer duplex exhibits several unusually shifted proton resonances which can be readily accounted for by the ring current contribution of the deoxyguanosyl and pyrenyl rings of the [BP]dG6 adduct. This solution structure of the (+)-trans-anti-[BP]dG.del 11-mer duplex where the pyrene ring intercalates into the helix with displacement of the modified deoxyguanosine into the major groove strikingly contrasts with our previous study on the (+)-trans-anti-[BP]dG.dC 11-mer duplex [Cosman, M., et al. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 1914-1918] where the benzo[a]pyrene ring is positioned in the minor groove without disruption of the Watson-Crick pairing at the [BP]dG.dC modification site. Thus, generation of the deletion site following removal of the dC opposite the (+)-trans-anti-[BP]dG results in a displacement of the entire [BP]dG residue toward the major groove and intercalation of the benzo[a]pyrene ring into the helix.

Fluoride is known to inhibit the photodynamic activity of aluminium phthalocyanine in a variety of biological systems. In order to gain insight into this phenomenon, the effect of fluoride on the photophysical properties of free and albumin-bound chloroaluminum phthalocyanine sulfonate (AlPcSn) were studied. The association constant of NaF with AlPcSn in aqueous solution was measured as 500 +/- 20 M-1. This binding affects the photophysical properties of the dye: the absorption bands in the visible range are blue-shifted by 6-8 nm, and this effect is mirrored in the fluorescence emission spectrum. Human serum albumin significantly quenched the dye fluorescence independent of the presence of fluoride ion. The transient absorption spectrum of the excited dye triplet is unchanged by NaF, but the quantum yield for its generation is increased by 50%, with no decrease in its lifetime. Formation of fluoroaluminum phthalocyanine complexes was also observed in tetrabutylammonium fluoride-assisted solutions in wet acetonitrile. The fluoro-AlPcSn complex is a better photosensitizer for generation of singlet oxygen than the original dye-hydroxyl ion complex, as confirmed using the imidazole-N,N-dimethyl-4-nitrosoaniline method. On the other hand, the fluoro-AlPcSn complex exhibits an intense inhibitory effect on photohemolysis of red blood cells (RBC) even after the cells are washed to remove free dye and fluoride prior to irradiation, indicating that once the dye is attached to the cellular site, the fluoride ligand is no longer prone to displacement (by hydroxyl ion, for example). Nonetheless, it is clear from the spectroscopic data that the new fluoro complex is an efficient sensitizer for photooxidation.(ABSTRACT TRUNCATED AT 250 WORDS)

A metabolite of benzo[a]pyrene, 9-r,10-t-dihydroxy-7,8-c-oxy-7,8,9,10- tetrahydrobenzo[a]pyrene (BPDE-III), that is not thought to be involved in carcinogenesis has nevertheless been shown to bind extensively to DNA in vitro. The adducts formed by this non-bay-region diol epoxide in Chinese hamster ovary cells are much less mutagenic than those formed by an isomeric diol epoxide that is carcinogenic. We have isolated and characterized three major adducts formed by in vitro reaction of BPDE-III with DNA. The major adduct, accounting for over half of the total is formed by reaction of BPDE-III with the N7 position of dGuo and is recovered after enzymatic digestion as an N7-Gua adduct. A second major adduct involves the N2 position of dGuo, while the third adduct is tentatively identified as a C8-substituted dGuo. Little or no reaction with deoxyadenosine residues is detected. The N7 adduct is unstable in DNA at 37 degrees C and is released as the modified base with a half-life of about 24 h. This adduct lability apparently leads to single-strand breaks and alkali-sensitive sites in the DNA and may account in part for some of the biological properties of BPDE-III adducts. This represents the first description of an N7-dGuo adduct that is formed in DNA as the major adduct by a diol epoxide derived from a carcinogenic polycyclic aromatic hydrocarbon.

Benzo[a]pyrene, an extremely potent procarcinogen and mutagen, is metabolized to a variety of products, including the ultimate carcinogen 7,8-dihydroxy-9,10-epoxy- 7,8,9,10-tetrahydrobenzo[a]pyrene. This product of biotransformation reacts with DNA, forming a series of adducts principally at the N2 position of guanine that differ in their stereochemistry and exhibit unique biological properties. In order to gain a better understanding of the effects on RNA synthesis of these adducts, we used purified bacteriophage T7 RNA polymerase to transcribe a series of templates containing one of four stereoisomerically pure BPDE-guanine lesions--(+)-trans-,(-)-trans-,(+)-cis-anti-N2-BPDE-guanine--or no damaged bases. To construct suitable double-stranded oligodeoxynucleotides for these studies, we annealed an 11-mer containing a site-specific stereoisomerically pure N2-BPDE-guanine adduct, a 37-mer, and a 10-mer to a complementary 58-base sequence of single-stranded DNA. The oligomers were ligated, purified, and reannealed. The resulting DNA template contained the promoter for T7 RNA polymerase and a BPDE adduct at position +16 following the transcription initiation site. The results of the transcription assays clearly demonstrate that each of the adducts inhibits elongation by T7 RNA polymerase, but they do so to significantly different extents, depending on the stereochemical characteristics of the BPDE-modified guanine. The order of inhibition is (+)-trans > (-)-trans > (+)-cis > (-)-cis, when the amount of full-length transcript for each is compared to that obtained for an unmodified template. Furthermore, premature termination of RNA synthesis occurs at or near the site of the BPDE lesion as evidenced by the formation of discrete, truncated transcripts. These results might be related to the fact that the pyrenyl moiety of the trans-BPDE adducts is situated in the minor groove of double-stranded DNA, but is quasi-intercalated into the double helix in the case of the cis stereoisomers.(ABSTRACT TRUNCATED AT 250 WORDS)

Benzo[c]phenanthrene diol epoxide can covalently bind to the exocyclic amino group of deoxyadenosine to generate [BPh]dA adducts where the polycyclic aromatic hydrocarbon is attached to the major groove edge of DNA. This paper reports on NMR-energy minimization structural studies of the (+)-trans-anti-[BPh]dA adduct positioned opposite dT in the sequence context d(C5-[BPh]A6-C7).d-(G16-T17-G18) at the 11-mer duplex level. The exchangeable and nonexchangeable protons of the benzo[c]phenanthrenyl moiety and the nucleic acid were assigned following analysis of two-dimensional NMR data sets in H2O and D2O solution. The solution structure of the (+)-trans-anti-[BPh]dA.dT 11-mer duplex has been determined by incorporating intramolecular and intermolecular proton-proton distances defined by upper and lower bounds deduced from NOESY data sets as restraints in energy minimization computations. The covalently attached benzo[c]phenanthrene ring intercalates to the 5'-side of the [BPh]-dA6 lesion site without disruption of the flanking Watson-Crick dC5.dG18 and [BPh]dA6.dT17 base pairs. The observed buckling of the intercalation cavity reflects the selective overlap of the intercalated phenanthrenyl ring with dT17 and dG18 bases on the unmodified strand. The structure provides new insights into how a polycyclic aromatic hydrocarbon covalently attached to the major groove edge of deoxyadenosine can still unidirectionally intercalate into the helix without disruption of the modified base pair. Our study establishes that among the contributing factors are a propeller-twisted [BPh]dA6.dT17 base pair, displacement of the carcinogen-DNA linkage bond from the plane of the dA6 base, the specific pucker adopted by the benzylic ring, and the propeller-like nonplanar geometry for the aromatic phenanthrenyl ring system. Our combined experimental-computational studies to date have now identified three structural motifs adopted by covalent polycyclic aromatic hydrocarbon-DNA adducts with their distribution determined by the chiral characteristics of individual stereoisomers and by whether the covalent adducts are generated at the minor or the major groove edge of the helix.

The deoxyribooligonucleotide 5'-d(CTCACATGTACACTCT) was reacted separately with the chiral diol epoxide isomers 7 beta, 8 alpha-dihydroxy-9 alpha, 10 alpha- epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene [(+)-anti-BPDE)] and 7 alpha, 8 beta-dihydroxy-9 beta, 10 beta-epoxy-7,8,9,10- tetrahydrobenzo[a]pyrene [(-)-anti-BPDE)], to produce the modified oligonucleotides 5'-d(CTCACATGBPDETACACTCT). Adducts in which either (+)-anti-BPDE or (-)-anti-BPDE are covalently bound via their C10 positions by trans addition to the exocyclic amino group of the single G residues were isolated and purified by HPLC methods. Snake venom phosphodiesterase (SVPD, phosphodiesterase I), which hydrolyzes DNA from the 3'-OH terminus to the 5'-end, digests the (+)-trans-anti-BPDE-oligonucleotide adducts at a significantly faster rate than that of the sterically different (-)-trans-anti-BPDE-oligonucleotide adducts. However, using spleen phosphodiesterase (SPD, phosphodiesterase II), which hydrolyzes DNA in the 5'-->3' direction, the opposite stereoselective resistance to digestion is observed. Using shorter BPDE-modified oligonucleotides as standards, the enzyme stall sites have been defined by gel electrophoresis methods; the most digestion-resistant phosphodiester linkage is the 5'-d(...T-G*...)-3' bond in the case of (+)-trans-BPDE-modified oligonucleotide adducts for both enzymes, SVPD and SPD (the starred G denotes the site of BPDE modification). In the case of the (-)-trans-BPDE-modified oligonucleotide adducts, the phosphodiester bond on the 3'-side of the modified G [5'-d(...G*-T...)-3'] is most resistant to digestion by both enzymes.(ABSTRACT TRUNCATED AT 250 WORDS)

Adducts derived from the covalent binding of two positional monomethyl-substituted isomers of a bay region chrysene diol epoxide to supercoiled pIBI30 DNA (2926 base pairs/genome) were prepared, and their characteristics were investigated by a combination of gel electrophoresis and flow linear dichroism techniques. The 5- and 6-methyl derivatives of trans-1,2-dihydroxy-anti-3,4-epoxy-1,2,3,4-tetrahydrochrysene [(+)-5- and (+)-6-MeCDE, respectively], both with 1R,2S,3S,4R stereochemistry, are characterized by significant differences in their biological activities [Melikian et al. (1988) Cancer Res. 48, 1781-1787]. When covalently bound to plasmid DNA, these two molecules give rise to striking differences in the gel electrophoretic and flow hydrodynamic characteristics of the modified supercoiled DNA. The hydrodynamic flow linear dichroism of linearized DNA molecules (obtained by EcoRI enzyme digestion of covalently closed supercoiled pIBI30 DNA), modified covalently with the highly tumorigenic and mutagenic (+)-5-MeCDE derivative, indicates that flexible joints, bends, or kinks are formed at the site of binding of (+)-5-MeCDE. Slab gel data, as well as ethidium bromide-titration tube agarose gel electrophoresis data, indicate that the formation of (+)-5-MeCDE-DNA lesions causes the removal of superhelical turns with an unwinding angle of 13 +/- 3 degrees per covalently bound polycyclic aromatic residue. In contrast, the biological inactive (+)-6-MeCDE does not significantly alter the characteristics of supercoiled DNA, the unwinding angle is only 2.7 +/- 1 degrees, and the changes in persistence lengths detected by the flow linear dichroism technique are significantly smaller than in the case of (+)-5-MeCDE-DNA adducts.(ABSTRACT TRUNCATED AT 250 WORDS)

The hydrophobic interactions of bulky polycyclic aromatic hydrocarbons with nucleic acid bases and the formation of noncovalent complexes with DNA are important in the expressions of the mutagenic and carcinogenic potentials of this class of compounds. The fluorescence of the polycyclic aromatic residues can be employed as a probe of these interactions. In this work, the interactions of the (+)-trans stereoisomer of the tetraol 7,8,9,10-tetrahydroxytetrahydrobenzo[a]pyrene (BPT), a hydrolysis product of a highly mutagenic and carcinogenic diol epoxide derivative of benzo[a]pyrene, were studied with 2'-deoxynucleosides in aqueous solution by fluorescence and UV spectroscopic techniques. Ground-state complexes between BPT and the purine derivatives 2'-deoxyguanosine (dG), 2'-deoxyadenosine (dA), and 2'-deoxyinosine (dI) are formed with association constants in the range of approximately 40-130 M(-1). Complex formation with the pyrimidine derivatives 2'-deoxythymidine (dT), 2'-deoxycytidine (dC), and 2'-deoxyuridine (dU) is significantly weaker. Whereas dG is a strong quencher of the fluorescence of BPT by both static and dynamic mechanisms (dynamic quenching rate constant k(DYN) = [2.5 +/- 0.4] x 10(9) M(-1)s(-1), which is close to the estimated diffusion-controlled value of approximately 5 x 10(9) M(-1)s(-1), both dA and dI are weak quenchers and form fluorescence-emitting complexes with BPT. The pyrimidine derivatives dC, dU, and dT are efficient dynamic fluorescence quenchers (k(DYN) approximately [1.5-3.0] x 10(9) M (-1)s(-1), with a small static quenching component due to complex formation evident only in the case of dT. None of the four nucleosides dG, dA, dC and dT are dynamic quenchers of BPT in the triplet excited state; the observed lower yields of triplets are attributed to the quenching of single excited states of BPT by 2'-deoxynucleosides without passing through the triplet manifold of BPT. Possible fluorescence quenching mechanisms involving photoinduced electron transfer are discussed. The strong quenching of the fluorescence of BPT by dG, dC and dT accounts for the low fluorescence yields of BPT-native DNA and of pyrene-DNA complexes.

Oligodeoxynucleotides modified site-specifically with dG-(+)-trans- and dG-(+)-cis-anti-BPDE (7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene) or dG-(-)-trans- and dG-(-)-cis-anti-BPDE were used as templates in primer extension reactions catalyzed by the Klenow fragment of Escherichia coli DNA polymerase I. The primer could be extended past the dG-(-)-trans-BPDE adduct with small amounts of dAMP incorporated opposite the lesion. A small amount of base deletions was also observed while, with the dG-(-)-cis-BPDE adduct, one- and two-base deletions predominated. When templates containing dG-(+)-trans-BPDE were used, small amounts of products containing one-base deletions were observed; with dG-(+)-cis-BPDE, substitution of dAMP opposite the lesion was also detected. The frequency of nucleotide insertion for dAMP opposite dG-(-)-trans-BPDE and the frequency of extension from the primer terminus containing the dA:dG-(-)-trans-BPDE pair were much higher than those observed with the other, stereochemically different BPDE adducts. Kinetic studies were in agreement with the results of the primer extension study. When the base flanking the 5' side of dG-BPDE was changed from dC to dT, the frequency of one-base deletions increased. We conclude that the trans- or cis-addition product of dG-(-)-anti-BPDE has a higher miscoding potential than dG-(+)-anti-BPDE in our model system and that G-->T transversions and deletions predominate. These observations are consistent with the types of mutations observed in vivo.

The fluorescence of mononucleoside adducts derived from the binding of anti-7 beta,8 alpha-dihydroxy-9 alpha,10 alpha-epoxy-7,8,9,10-tetrahydrobenzo [a]pyrene (BPDE I) to N6-deoxyadenosine (BPDE-dA adducts) is 10-100 times stronger (depending on the methanol/water solvent composition) than the fluorescence of adducts derived from the binding of this diol epoxide derivative to N2-deoxyguanosine. It is shown here that these fluorescence characteristics can be used to quantitate the relatively low yields of BPDE-dA adducts by fluorescence detection when BPDE-modified DNA is subjected to enzymatic degradation to the mononucleoside levels, followed by HPLC analysis of the digests.

This paper reports on the solution structure of the (+)-cis-anti-[BP]dG adduct positioned opposite dC in a DNA oligomer duplex which provides the first experimentally based solution structure of an intercalative complex of a polycyclic aromatic hydrocarbon covalently bound to the N2 of deoxyguanosine. The combined NMR-energy minimization computation studies were undertaken on the (+)-cis-anti-[BP]dG adduct embedded in the same d(C5-[BP]G6-C7).d(G16-C17-G18) trinucleotide segment of the complementary 11-mer duplex studied previously with the stereoisomeric trans adducts. The exchangeable and nonexchangeable protons of the benzo[a]pyrenyl moiety and the nucleic acid were assigned following analysis of two-dimensional NMR data sets in H2O and D2O solution. The solution structure of the (+)-cis-anti-[BP]dG-dC 11-mer duplex has been determined by incorporating intramolecular and intermolecular proton-proton distances defined by upper and lower bounds deduced from NOESY data sets as restraints in energy minimization computations. The benzo[a]pyrene ring of [BP]dG6 is intercalated between intact Watson-Crick dC5.dG18 and dC7.dG16 base pairs in a right-handed DNA helix. The benzylic ring is in the minor groove while the pyrenyl ring sacks with flanking dC5 and dC7 bases on the same strand. The deoxyguanosine ring of [BP]dG6 is not Watson-Crick base paired but displaced into the minor groove with its plane parallel to the helix axis and stacks over the sugar ring of dC5. The dC17 base on the partner strand is displaced from the center of the helix toward the major groove by the intercalated benzo[a]pyrene ring. This intercalative structure of the (+)-cis-anti-[BP]dG-dC 11-mer duplex exhibits several unusually shifted proton resonances which can be readily accounted for by the ring current contributions of the deoxyguanosine and pyrenyl rings of the [BP]dG6 adduct. Several phosphorus resonances are shifted to low and high field of the unperturbed phosphorus spectral region and have been assigned to internucleotide phosphates centered about the [BP]dG6 modification site. These studies define the changes in the helix at the central trinucleotide segment needed to generate the intercalation site for the covalently bound (+)-cis-anti-[BP]dG adduct.(ABSTRACT TRUNCATED AT 400 WORDS)

The base-sequence dependence of the yield of the covalent binding reaction of (+)-anti-7 beta, 8 alpha-dihydroxy-9 alpha, 10 alpha-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene [(+)-anti-BPDE] with the exocyclic amino group of guanine surrounded by different flanking bases X and Y in the single-stranded oligonucleotide d(CTATXGYTATC) was investigated. With an initial ratio of [(+)-anti-BPDE]/[oligonucleotide strand] = 2, the percentage of modified strands varied from 20 +/- 2% when the modified dG was surrounded by pyrimidines to 5-7% when the central dG was surrounded by purines. The trans/cis ratio of (+)-anti-BPDE-N2-dG adducts was in the range of 3-5. The lower reaction yields observed when the modified guanine residues in single-stranded oligonucleotides are surrounded by purines rather than by pyrimidines is tentatively attributed (1) to steric effects arising from the presence of the bulkier purines flanking the reacting dG moieties on the 5'- and 3'-sides and/or (2) to noncovalent interactions between anti-BPDE and neighboring purines which decrease the probability of optimal alignment for covalent binding between the interacting moieties in the bimolecular transition-state complex. Noncovalent intercalation of (+)-anti-BPDE prior to the covalent binding reaction is not a relevant process in the case of single-stranded oligonucleotides and is therefore not a critical requirement for obtaining high yields of covalent trans- and cis-(+)-anti-BPDE-N2-dG adducts in these oligonucleotide sequences.

The unwinding of supercoiled phi X174 RFI DNA induced by the tumorigenic (+) and non-tumorigenic (-) enantiomers of trans-7,8-dihydroxy-anti-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) has been investigated by agarose slab-gel and ethidium titration tube gel electrophoresis. The differences in adduct conformations were verified by flow linear dichroism techniques. Both enantiomers cause a reversible unwinding by the formation of noncovalent intercalative complexes. The effects of covalently bound BPDE residues on the electrophoretic mobilities of the RF I DNA form in agarose gels were investigated in detail in the range of binding ratios rb approximately 0.0-0.06 (covalently bound BPDE residues/nucleotide). In this range of rb values, there is a striking difference in the mobilities of (+)-BPDE- and (-)-BPDE-adducted phi X174 DNA in agarose slab-gels, the covalently bound (+)-BPDE residues causing a significantly greater retardation than (-)-BPDE residues. Increasing the level of covalent adducts beyond rb approximately 0.06 in the case of the (+)-BPDE enantiomer, leads to further unwinding and a minimum in the mobilities (corresponding to comigration of the nicked form and the covalently closed relaxed modified form) at rb 0.10 +/- 0.01; at still higher rb values, rewinding of the modified DNA in the opposite sense is observed. From the minimum in the mobility, a mean unwinding angle (per BPDE residue) of theta = 12 +/- 1.5 degrees is determined, which is in good agreement the value of theta = 11 +/- 1.8 degrees obtained by the tube gel titration method. Using this latter method, values of theta = 6.8 +/- 1.7 degrees for (-)-BPDE-phi X174 adducts are observed. It is concluded that agarose slab gel techniques are not suitable for determining unwinding angles for (-)-BPDE-modified phi X174 DNA because the alterations in the tertiary structures for rb < 0.06 are too small to cause sufficiently large changes in the electrophoretic mobilities. The major trans (+)-BPDE-N2-guanosine covalent adduct is situated at external binding sites and the mechanisms of unwinding are therefore different from those relevant to noncovalent intercalative BPDE-DNA complexes or to classical intercalating drug molecules; a flexible hinge joint and a widening of the minor groove at the site of the lesion may account for the observed unwinding effects. The more heterogeneous (-)-BPDE-nucleoside adducts (involving cis and trans N2-guanosine, and adenosine adducts) are less effective in causing unwinding of supercoiled DNA for reasons which remain to be elucidated.

Adducts derived from the reaction of the benzo[a]pyrene metabolite model compound (+)-anti-7 beta,8 alpha-dihydroxy-9 alpha,10 alpha-epoxy-7,8,9, 10-tetrahydrobenzo[a]pyrene [(+)-BPDE] with the single-stranded oligodeoxynucleotide 5'-d(TATGCGTAT) were obtained according to direct synthesis techniques described earlier [Cosman, M., Ibanez, V., Geacintov, N. E., and Harvey, R. G. (1990) Carcinogenesis 11, 1667-1672]. Four major adducts, involving trans and cis addition (trans/cis adduct ratio approximately 4.5) of (+)-BPDE to the exocyclic amino groups of guanines G4 and G6 (the numbers denote the positions of the guanines counted from the 5'-side) were obtained. These adducts can be separated from one another by reverse-phase high-performance liquid chromatography methods. The site of BPDE binding on either G4 or G6 can be determined from the electrophoresis band patterns on 20% polyacrylamide gels of the BPDE-modified oligonucleotides subjected to the G+A and G Maxam-Gilbert strand cleavage reactions [Maxam, A. M., and Gilbert, W. (1980) Methods. Enzymol. 65, 499-560]. The electrophoresis gel band patterns are different for unmodified DNA and the two different BPDE-modified oligonucleotides because (1) the strand cleavage fragments bearing BPDE residues migrate slower than the corresponding fragments derived from the unmodified oligonucleotide and (2) strand cleavage tends to be inhibited on the 5'-sides of BPDE-modified guanines in the G+A, but not the G reaction.