1H-NMR Assignments of the Non-exchangeable Protons of the Consensus Donor Exon:Intron Junction d(CpApGpGpTpApApGpT)

Abstract

The consensus donor exon:intron junction d(CpApGpGpTpApApGpT) has been synthesized by a modified phosphotriester method. The non-self-complementary nonamer has, in principle, only two G,C or four A,T points of self-recognition. The inference that it exists in the single strand form at 20°C was confirmed by temperature variable ¹H-NMR and NOE measurements. The proton assignments were secured using two-dimensional COSY which provided intra-nucleotide correlations, then NOE difference measurements as well as inversion recovery T₁ experiments. Systematic procedures were developed for the assignment of the individual bases and their component protons based on the effects of molecular environment on chemical shifts. These latter procedures should be useful for the assignment of other random-coil single strand oligodeoxyribonucleotides.     

Structure and Conformation of the Duplex Consensus Acceptor Exon: Intron Junction d [(CpTpApCpApGpGpT) (ApCpCpTpGpTpApG)] deduced from High-Field 1H-NMR of Non-Exchangeable and Imino Protons

Abstract

The complementary consensus acceptor exon:intron junction d(ApCpCpTpGpTpApG) has been synthesized by a modified phosphotriester method. The non self-complementary octamer exists in the random coil form in aqueous buffer at 20°C as evidenced by temperature variable ¹H-NMR and NOE measurements. The non-exchangeable proton assignments were secured using a combination of techniques including two-dimensional COSY, NOESY and ¹H-¹H-INADEQUATE. The octamer was annealed with the primary consensus sequence d(CpTpApCpApGpGpT). Confirmation of complete duplex formation was confirmed by detection and assignment of imino protons in D₂O:H₂O mixtures. Assignment of the nonexchangeable proton signals in the duplex consensus junction was then secured by a combination of two-dimensional COSY correlations, NOESY and NOE experiments. Determination of individual vicinal coupling constants in the component deoxyribose moieties permitted deduction of the population of S conformations in this sequence. It is concluded that the consensus acceptor junction exists in solution in a conformation belonging to the B family, and that the bases are oriented anti. In addition the deoxyribose moieties in the 5′ regions exist predominantly in the S form (2′endo—3′exo) whereas those residues on or adjacent to the junction on the primary strand show more N character (2′exo—3′endo). The contiguous bases A₅-G₆ (adjacent to the junction) and A₁₅-G₁₆ are stacked more closely than the other neighbor bases in this duplex sequence. These subtle structural and conformational differences in the exon:intron junction may serve as recognition signals for these critical sites in the genome.     

1H and 31P-NMR assignments of the non-exchangeable protons of the consensus acceptor exon:intron junction d(CpTpApCpApGpGpT)

The consensus acceptor exon:intron junction d(CpTpApCpApGpGpT) has been synthesized by a modified phosphotriester method. The non-self complementary octamer exists in the single strand form in aqueous buffer at 20 degrees C as evidenced by temperature variable 1H-NMR and NOE measurements. The non-exchangeable proton assignments were secured using a combination of techniques including two-dimensional COSY, NOESY and the double quantum technique 1H-1H-INADEQUATE as well as inversion recovery T1 experiments. The new technique of 31P-1H shift correlation is particularly valuable in removing certain ambiguities in the sugar proton assignments. Characteristic chemical shifts for the base protons which are determined by their immediate molecular environments are also useful in assignments. The consensus acceptor exon:intron junction adopts a random coil conformation in solution under the experimental conditions employed.     

1H-NMR Assignment of Single Strand and Duplex Splice Domain of the Consensus Donor Exon:Intron Junction

Abstract

The high field ¹H-NMR assignments of a single strand consensus donor exon: intron junction and that of the duplex splice domain has been achieved using 2D-NMR and additional techniques.     

Bisintercalation of ditercalinium into a d(CpGpApTpCpG)2 minihelix: a 1H- and 31P-NMR study

The structure of the complex formed in aqueous solution between ditercalinium, a bisintercalating drug, and the self-complementary hexanucleotide d(CpGpApTpCpG)2 is investigated by 400-MHz 1H-nmr and 162-MHz 31P-nmr. Whatever the drug to helix ratio, ditercalinium occurred in the bound form, whereas free and complexed hexanucleotide are in slow exchange. This allows unambiguous resonance assignment through two-dimensional chemical exchange experiments. The strong upfield shifts measured on most aromatic protons on both drug and bases as well as on DNA imino protons are consistent with bisintercalation of the dimer. Nuclear Overhauser effects observed between drug and nucleotide protons give a defined geometry for complexation, and suggest a DNA conformational change upon drug binding.     

The 31P-NMR spectrum of the dodecamer d(GACGATATCGTC).

The resonances in the 31P-NMR spectrum of the dodecamer d(GACGATATCGTC) have been assigned by regiospecific labelling with oxygen-17. All 11 resonances are clearly resolved at 26 degrees C. Most noticeably, individual resonances of the dinucleoside phosphates d(CpG), d(TpC), d(GpA) and d(ApT) which occur more than once can clearly be distinguished. This indicates that the position of the phosphate group in the oligomer influences its 31P-NMR shift. This observation is in agreement with what has been found for the 31P-NMR spectra of d(CGCGAATTCGCG) [Ott, J. and Eckstein, F. (1985) Biochemistry 24] and d(GGAATTCC) [Connolly, B.A. and Eckstein, F. (1984) Biochemistry 23, 5523-5527]. In general, the chemical shift appears the more at higher field the more central the dinucleoside phosphate is located in the oligomer. Exceptions are the resonances of dinucleoside phosphates of the type 5'-PyPu-3' which appear at lower field than expected from this rule. A reasonable correlation between 31P-NMR chemical shifts and the sum function of the base plane roll angles derived from Calladine's rule [Calladine, C.R. (1982) J. Mol. Biol. 161, 343-352] exists.     

1H- and 31P-NMR studies of ditercalinium binding to a d(GCGC)2 and d(CCTATAGG)2 minihelices: a sequence specificity study

The structures of the complexes formed in aqueous solution between ditercalinium, a bis-intercalating drug, and both the self-complementary tetranucleotide d(GCGC)2 and octanucleotide d(CCTATAGG)2, have been investigated by 400-MHz 1H-nmr and 162-MHz 31P-nmr. All the nonexchangeable protons, as well as the exchangeable imino protons and the phosphorus signals, have been assigned. Both oligonucleotides have been shown to adopt a right-handed B-DNA type structure. The addition of ditercalinium to the oligonucleotides lead to the formation of complexes in slow exchange at the nmr time scale with the free helices. At all drug-to-helix ratios studied, the ditercalinium was found in the bound form, whereas free and complexed oligonucleotides were in slow exchange, allowing resonance assignments through two-dimensional chemical exchange experiments. for d(GCGC)2 the strong upfield shifts induced on all aromatic protons of both the bases and the drug by complexation with ditercalinium suggest an interaction by intercalation of the two rings. However, the loss of twofold symmetry upon binding, as well as the chemical shift variation of the drug proton signals of one of the chromophores with temperature and concentration, favor a model in which the drug-nucleotide complexes have one ring of the drug intercalated and the other stacked on top of the external base pair. The intermolecular contacts between drug protons and nucleotide protons give a defined geometry for complexation that is consistent with the proposed model. In contrast, with d(CCTATAGG)2 several drug-nucleotide complexes were formed and a large increase in line broadening was observed at high drug-to-DNA ratios, precluding a detailed analysis of these complexes. However, the large upfield shift in the imino proton resonances together with the shielding of the ditercalinium ring protons favor a model with bis-intercalation of ditercalinium. This model is supported by the downfield shift of at least 4 out of 14 phosphorus signals. The results are compared with those obtained on ditercalinium binding to the homologous sequences d(CGCG)2 and d(TTCGCGAA)2, and discussed in terms of sequence specificity.     

Comparison of the bis-intercalating complexes formed between either ditercalinium or a flexible analogue and d(CpGpCpG)2 or d(TpTpCpGpCpGpApA)2 minihelices: 1H- and 31P-NMR analyses.

The 400-MHz 1H- and 162-MHz 31P-nmr have been used to study complexes constituted by (a) the d(TpTpCpGpCpGpApA)2 or the d(CpGpCpG)2 self-complementary oligonucleotides and (b) two bifunctional 7H-pyrido [4,3-c] carbazole dimer drugs, the antitumoral ditercalinium (NSC 366241), a dimer with a rigid bis-piperidine linking chain and its pharmacologically inactive analogue, a dimer with a flexible spermine-like linking chain. Nearly all proton and phosphorus signals have been assigned by two-dimensional (2D) nmr (correlated spectroscopy, homonuclear Hartmann-Hahn, nuclear Overhauser enhancement spectroscopy, 2D 31P (1H) heteronuclear correlated spectroscopy and 31P-31P chemical exchange experiments). Both drugs bis-intercalate into the two CpG sites. The complexes show small differences in the position of the 7H-pyrido [4,3-c] carbazole ring into the intercalation site and possibly in the ribose-phosphate backbone deformation. However, the inactive analogue exhibits a longer residence lifetime in octanucleotide than the ditercalinium does. All these results are discussed in terms of differences in dimer activities.     

Phosphocholine and phosphoethanolamine during chick embryo myogenesis: a (31)P-NMR study

Elevated contents of phosphoethanolamine (Etn-P) and/or phosphocholine (Cho-P), a common feature of most tumours with respect to normal counterparts, may also occur in non-cancerous proliferating tissues. The significance of these alterations in relation to cell proliferation, differentiation and maturation is scarcely understood. In this work, the Cho-P and Etn-P pools were measured by (31)P-NMR in extracts of chick embryo pectoral muscle at different days of development. The average concentration of these metabolites exhibited the highest values (respectively, 1.5 and 3.0 micromol/mg DNA) on days 9-11 and decreased at later stages of myogenesis. While, however, Cho-P maintained substantial levels (above 1.0 micromol/mg DNA) also during myotube formation (days 11-18) and stepwise decreased (to about 0.5 micromol/mg DNA) upon fibres' maturation, Etn-P gradually decreased between day 11 and hatching time (down to about 0.2 micromol/mg DNA). These results demonstrate that significant changes may occur in the steady-state pools of these metabolites during normal in vivo cellular development and differentiation, and are consistent with: (a) high rates of phospholipid biosynthesis reported in the literature for proliferating myoblasts; (b) sustained phosphatidylcholine synthesis maintained also during myoblast fusion; and (c) decreased requirement of phospholipid synthesis in the last phase of in ovo myofibre maturation.     

Stopped-flow kinetic, 1H, and 31P NMR analysis of the intercalation of ethidium with poly d(G-C) X d(G-C)

In a low salt buffer (0.011 M Na+) stopped-flow kinetic results for the SDS driven dissociation of an ethidium-Poly d(G-C) X d(G-C) complex are 8.7, 23, and 58.5 s-1 at 20, 30, and 40 degrees C, respectively. These results predict that in NMR experiments at high field strengths, ethidium should be in slow exchange among polymer binding sites. This has been found to be the case for both 31P (109 MHz) and 1H (imino proton spectra in H2O at 270 MHz) experiments. At higher salt, and/or higher temperature, and/or lower field, the bound and free peaks are no longer resolved in the NMR spectra. Good agreement is obtained between the stopped-flow kinetic results and the coalescence temperature observed in NMR experiments. Imino protons in base pairs on both sides of the intercalated ethidium are shifted approximately one ppm upfield while only the phosphate groups at the intercalation site experience large chemical shifts.     

1H-NMR and (31)P-NMR analysis of energy metabolism of quiescent and motile turbot (Psetta maxima) spermatozoa

31P-NMR and (1)H-NMR were used to monitor changes of several compounds with high-energy bonds and metabolites prior to and after the initiation of motility of turbot spermatozoa (Psetta maxima). The obtained (31)P-NMR spectra revealed the presence of phosphomonoesters, phosphodiester, intracellular inorganic phosphate (Pi), phosphocreatine (PCr), and free nucleotide triphosphate. Following the activation of motility, the di- and tri-phosphate nucleotides, PCr, phosphomonoesters levels dropped while Pi levels increased. A significant increase of lactate was also seen at the end of the swimming phase. The compositions of seminal fluid and urine were also determined. Lipoproteins, formic acid, amino acid, and citric acid were detected in seminal fluid. Dimethyl amine, trimethylamine, and trimethylamine oxyde were found in urine. These data suggest that at least a part of the energy required during the swimming phase results from anaerobic fermentation and oxidative phosphorylation. J. Exp. Zool. 286:513-522, 2000.     

Interaction between beta-Purothionin and dimyristoylphosphatidylglycerol: a (31)P-NMR and infrared spectroscopic study

The interaction of beta-purothionin, a small basic and antimicrobial protein from the endosperm of wheat seeds, with multilamellar vesicles of dimyristoylphosphatidylglycerol (DMPG) was investigated by (31)P solid-state NMR and infrared spectroscopy. NMR was used to study the organization and dynamics of DMPG in the absence and presence of beta-purothionin. The results indicate that beta-purothionin does not induce the formation of nonlamellar phases in DMPG. Two-dimensional exchange spectroscopy shows that beta-purothionin decreases the lateral diffusion of DMPG in the fluid phase. Infrared spectroscopy was used to investigate the perturbations, induced by beta-purothionin, of the polar and nonpolar regions of the phospholipid bilayers. At low concentration of beta-purothionin, the temperature of the gel-to-fluid phase transition of DMPG increases from 24 degrees C to ~33 degrees C, in agreement with the formation of electrostatic interactions between the cationic protein and the anionic phospholipid. At higher protein concentration, the lipid transition is slightly shifted toward lower temperature and a second transition is observed below 20 degrees C, suggesting an insertion of the protein in the hydrophobic core of the lipid bilayer. The results also suggest that the presence of beta-purothionin significantly modifies the lipid packing at the surface of the bilayer to increase the accessibility of water molecules in the interfacial region. Finally, orientation measurements indicate that the alpha-helices and the beta-sheet of beta-purothionin have tilt angles of ~60 degrees and 30 degrees, respectively, relative to the normal of the ATR crystal.     

Nucleotide chain length and the morphology of complexes with cationic amphiphiles: (31)P-NMR observations

31P-NMR and UV spectroscopies were used to study the interactions between cationic amphiphile-containing lipid bilayers and either a phosphorothioate oligonucleotide (OligoS) (n=21) or polyadenylic acid (PolyA) (n approximately 18,000). Multilamellar vesicles (MLVs) were composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) or 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) in binary mixture with either of the cationic lipids, N-[1-(2, 3-dioleoyloxy)propyl]-N',N',N'-trimethylammonium chloride (DOTAP) or cetyltrimethylammonium bromide (CTAB). A UV-difference assay showed that OligoS binding ceased above a 1:1 anion/cation ratio, while PolyA binding continued until a 2:1 ratio was reached, indicating a 'flat' conformation for bound OligoS, but not necessarily for PolyA. Cross-polarization (31)P-NMR of the nucleotide chains bound to 100% DOTAP MLVs produced spectra virtually identical to those of dry powders of OligoS or PolyA, indicating effective immobilization of the surface-bound nucleotide chains. Hahn echo (31)P-NMR showed that MLVs composed of binary mixtures of POPC with DOTAP or CTAB retained a lamellar bilayer architecture upon adding nucleotide chains. At less than stoichiometric anion/cation ratios little or no signal attributable to free nucleotide chains was visible. A narrow signal at the chemical shift expected for phosphorothiodiesters or phosphodiesters became visible at greater levels of added OligoS or PolyA, respectively, indicating the presence of mobile nucleotide chains. Salt addition caused complete desorption of the nucleotide chains. When POPC was replaced with DOPE, binding of OligoS or PolyA produced non-bilayer lipid phases in the presence of DOTAP, but not in the presence of CTAB.     

Reassessment of structural characteristics of the d(CGCG)2:actinomycin D complex from complete 1H and 31P NMR

Complexes formed between Actinomycin D (ActD) and the tetranucleotides d(AGCT)2 and d(CGCG)2 were studied in detail by one and two-dimensional 1H and 31P NMR. The 31P two dimensional chemical exchange experiment, at room temperature on saturated complexes (1:1), showed unambiguously that the asymmetrical phenoxazone ring binds to the unique GC site under the two possible orientations in the d(AGCT)2 tetranucleotide but adopts a single orientation in the d(CGCG)2 tetranucleotide. For the d(CGCG)2:Act D saturated complex, complete assignments of all protons and phosphorus signals of the two-nucleotide strands, as well as of the two cyclic pentapeptide chains has allowed us to study in details the conformational features of the complex from NOE and coupling constants analysis. The tetranucleotide remains in a right-handed duplex, but the sugar puckers are modified for residues at the intercalation site. A uniform C2' endo pucker is observed for residues on the strand facing the quinoid side of the phenoxazone ring while a C2' endo-C3-endo equilibrium about 60% of C2' endo is proposed for the two residues on the strand facing the benzenoid side of the phenoxazone ring. In contrast to previous studies on ActD-DNA interactions, we have been able to measure the 3J phosphorus-proton coupling constants at the intercalation site but also adjacent to it, showing that 31P chemical shifts are not simply related to the backbone conformation. Molecular mechanics calculations, using empirical distances deduced from NOE effects as restrained distances during minimizations, led to a model differing mainly from those previously published by orientation of the N methyl groups of both N-Methyl-Valines.     

A 31P-NMR spin-lattice relaxation and 31P[1H] nuclear Overhauser effect study of sonicated small unilamellar phosphatidylcholine vesicles.

The motional properties of the inner and outer monolayer headgroups of egg phosphatidylcholine (PC) small unilamellar vesicles (SUV) were investigated by 31P-NMR temperature-dependent spin-lattice relaxation time constant (T1) and 31P[1H] nuclear Overhauser effect (NOE) analyses. Three different aspects of the dynamics of PC headgroups were investigated using the T1 analysis. First, differences in the dynamics of the headgroup region of both surfaces of the SUV were measured after application of a chemical shift reagent, PrCl3, to either the extra- or intravesicular volumes. Second, the ability of the T1 experiment to resolve the different motional states was evaluated in the absence of shift reagent. Third, comparison between correlation times obtained from a resonance frequency dependent 31P[1H] NOE analysis allowed a determination of the applicability of a simplified motional model to describe phosphorus dipolar relaxation. Temperature-dependent 31P-NMR T1 values obtained for the individual monolayers at 81.0 and 162.0 MHz were modelled assuming that phosphorus undergoes both a dipolar and an anisotropic chemical shielding relaxation mechanism, each being described by the same correlation time, tau. At 162.0 MHz, the position of the T1 minimum for the inner monolayer was 9 degrees higher than that of the outer region, indicating a higher level of motional restriction for the inner leaflet, in agreement with 31P[1H] NOE measurements. The 162.0 MHz T1 profile of the combined SUV monolayers exhibited a smooth minimum located at the midpoint of the monolayer minima positions, effectively masking the presence of the individual surfaces. 31P[1H] NOE results obtained at 32.3, 81.0 and 162.0 MHz did not agree with those predicted from a simple dipolar relaxation model. These results suggest a T1-temperature method can neither discriminate two or more closely related motional time scales in a heterogeneous environment (such as incorporation of protein into lipid bilayers) nor allow accurate determination of the correlation time at the position of the minimum when the dipolar relaxation rate makes a significant contribution to the overall rate.     

d(TTTCCTCGCCGGAAA)的一维NMR氢谱分析

单链d(TTTCCTCGCCGGAAA)易溶于水,且由于其本身存在序列特异性,即可形成分子内“发夹”结构,本实验分别测得其在全重水(D2O)、92?O 8%H2O溶液中的一维1H谱,认为环出区域碱基质子的共振峰与其他同种质子的共振峰有明显的区别,主要表现在其共振峰会明显移向高场区。     

Critical oxygen tension in rat brain: a combined (31)P-NMR and EPR oximetry study

The relationship between cerebral interstitial oxygen tension (Pt(O(2))) and cellular energetics was investigated in mechanically ventilated, anesthetized rats during progressive acute hypoxia to determine whether there is a "critical" brain Pt(O(2)) for maintaining steady-state aerobic metabolism. Cerebral Pt(O(2)), measured by electron paramagnetic resonance oximetry, decreased proportionately to inspired oxygen fraction. (31)P-nuclear magnetic resonance measurements revealed no changes in P(i), phosphocreatine (PCr)/P(i) ratio, or intracellular pH when arterial blood oxygen tension (Pa(O(2))) was reduced from 145.1 +/- 11.7 to 56.5 +/- 4.4 mmHg (means +/- SE). Intracellular acidosis, a sharp rise in P(i), and a decline in the PCr/P(i) ratio developed when Pa(O(2)) was reduced further to 40.7 +/- 2.3 mmHg. The corresponding Pt(O(2)) values were 15.1 +/- 1.8, 8.8 +/- 0.4, and 6.8 +/- 0.3 mmHg. We conclude that over a range of decreasing oxygen tensions, cerebral oxidative metabolism is not sensitive to oxygen concentration. Oxygen becomes a regulatory substrate, however, when Pt(O(2)) is decreased to a critical level.     

Lipid analysis of human HDL and LDL by MALDI-TOF mass spectrometry and (31)P-NMR

The analysis of HDL and LDL is important for the further understanding of atherosclerosis because changes of the protein and lipid moieties occur under pathological conditions. Because destruction of lipids leads to the formation of well-defined products such as lysophospholipids or chlorohydrins, methods that allow their fast and reliable determination would be useful. In this study, matrix-assisted laser desorption and ionization time-of-flight mass spectrometry (MALDI-TOF MS) was applied for the analysis of the lipid composition of human lipoproteins. These data were compared with high resolution (31)P-NMR spectroscopy. Differences between LDL and HDL in sphingomyelin and phosphatidylcholine content could be monitored by NMR and mass spectrometry, and differences with respect to the extraction efficiency were found by MALDI-TOF MS. Additionally, treatment of LDL with hypochlorite and phospholipase A(2) resulted in marked changes (formation of chlorohydrines and lysolipids). Lysophosphatidylcholines were detectable by both methods, whereas MALDI-TOF MS failed to detect chlorohydrines of phospholipids.We conclude that MALDI-TOF MS provides rapidly a reliable lipid profile of lipoproteins. However, a previous lipid separation must be performed to detect lipid oxidation products. NMR can be directly applied, but suffers from lower sensitivity, and provides only limited information on fatty acid composition.     

Intercalation of aflatoxin B1 in two oligodeoxynucleotide adducts: comparative 1H NMR analysis of d(ATCAFBGAT).d(ATCGAT) and d(ATAFBGCAT)2

8,9-Dihydro-8-(N7-guanyl-[d(ATCGAT)])-9-hydroxyaflatoxin B1.d(ATCGAT) and 8,9-dihydro-8-(N7-guanyl-[d(ATGCAT)])-9-hydroxyaflatoxin B1.8,9-dihydro-8-(N7-guanyl-[d(ATGCAT)])-9-hydroxyaflatoxin B1 were prepared by direct addition of afltoxin B1 8,9-epoxide to d(ATCGAT)2 and d(ATGCAT)2, respectively. In contrast to reaction of aflatoxin B1 8,9-epoxide with d(ATCGAT)2 which exhibits a limiting stoichiometry of 1:1 aflatoxin B1:d(ATCGAT)2 [Gopalakrishnan, S., Stone, M. P., & Harris, T. M. (1989) J. Am. Chem. Soc. 111, 7232-7239], reaction of aflatoxin B1 8,9-epoxide with d(ATGCAT)2 exhibits a limiting stoichiometry of 2:1 aflatoxin B1:d(ATGCAT)2. 1H NOE experiments, nonselective 1H T1 relaxation measurements, and 1H chemical shift perturbations demonstrate that in both modified oligodeoxynucleotides the aflatoxin moiety is intercalated above the 5'-face of the modified guanine. The oligodeoxynucleotides remain right-handed, and perturbation of the B-DNA structure is localized adjacent to the adducted guanine. Aflatoxin-oligodeoxynucleotide 1H NOEs are observed between aflatoxin and the 5'-neighbor base pair and include both the major groove and the minor groove. The aflatoxin methoxy and cyclopentenone ring protons face into the minor groove; the furofuran ring protons face into the major groove. No NOE is observed between the imino proton of the modified base pair and the imino proton of the 5'-neighbor base pair; sequential NOEs between nucleotide base and deoxyribose protons are interrupted in both oligodeoxynucleotide strands on the 5'-side of the modified guanine. The protons at C8 and C9 of the aflatoxin terminal furan ring exhibit slower spin-lattice relaxation as compared to other oligodeoxynucleotide protons, which supports the conclusion that they face into the major groove. Increased shielding is observed for aflatoxin protons; chemical shift perturbations of the oligodeoxynucleotide protons are confined to the immediate vicinity of the adducted base pair. The imidazole proton of the modified guanine exchanges with water and is observed at 9.75 ppm. The difference in reaction stoichiometry is consistent with an intercalated transition-state complex between aflatoxin B1 8,9-epoxide and B-DNA. Insertion of aflatoxin B1-8,9 epoxide above the 5'-face of guanine in d(ATCGAT)2 would prevent the binding of a second molecule of aflatoxin B1 8,9-epoxide. In contrast, two intercalation sites would be available with d(ATGCAT)2.(ABSTRACT TRUNCATED AT 400 WORDS)