Duplex Adduct Research Articles

Observation of a Single-Stranded DNA/Pyrrolobenzodiazepine Adduct

Pyrrolobenzodiazepine (PBD) antitumor agents have, to date, only been observed to bind to duplex DNA, apparently requiring a minor groove environment for covalent bond formation between their C11-position and the C2-NH(2) functionality of a guanine base. Using an HPLC/MS assay we have now observed and isolated for the first time PBD adducts with single-stranded DNA fragments. Surprisingly, these adducts could only be formed through dissociation of duplex DNA adducts and not by direct interaction of PBDs with single-stranded DNA. They were sufficiently stable for characterization by MALDI-TOF-MS and remained intact after storing at -20 °C for at least 20 days, although the PBD became detached from the DNA within 7 days if stored at room temperature. Furthermore, addition of a complementary strand allowed the duplex adduct to reform. The relative stability of single-stranded PBD/DNA adducts despite a complete loss of minor groove structure was further confirmed by CD spectroscopic analysis. The CD signal induced by the presence of a PBD molecule in the single-stranded adducts remained prominent despite heating for 2 h at 50-60 °C, thus indicating their relatively robust nature.

Nuclear Magnetic Resonance Solution Structures of Inter- and Intrastrand Adducts of DNA Cross-Linker SJG-136

SJG-136 (1) is a sequence-selective DNA-interactive agent that is about to enter phase II clinical trials for the treatment of malignant disease. Previous studies on the pyrrolo[2,1-c][1,4]benzodiazepine (PBD) dimers, typified by SJG-136 and DSB-120 (2), have shown that these planar ligands react with the exocyclic NH(2) groups of two guanine bases in the base of the minor groove of DNA to form an irreversible interstrand cross-linked sequence-specific adduct. Using high-field NMR, we have characterized and modeled the previously predicted interstrand duplex adduct formed by SJG-136 with the self-complementary 5'-d(CICGATCICG)(2) duplex (4). This first SJG-136 NMR-refined adduct structure has been compared with previous high-field NMR studies of the adducts of the closely related PBD dimer DSB-120 with the same duplex and of the adduct of tomaymycin (3) formed with 5'-d(ATGCAT)(2). Surprisingly, the SJG-136 duplex adduct appears to be more closely related to the tomaymycin adduct than to the DSB-120 adduct with respect of the orientation and depth of insertion of the ligand within the minor groove. The intrastrand duplex adduct formed in the reaction of SJG-136 with the noncomplementary 5'-d(CTCATCAC)·(GTGATGAG) duplex (5) has also been synthesized and modeled. In this duplex adduct, the nature of the cross-link was confirmed, the central guanines were identified as the sites of alkylation, and the stereochemical configuration at C11 at both ends of the SJG-136 molecule was determined to be S. The NMR-refined solution structures produced for the intrastrand adduct confirm the previously proposed structure (which was based solely on mass spectroscopy). Both the inter- and intrastrand SJG-136 duplex adducts form with minimal distortion of the DNA duplex. These observations have an impact on the proposal for the mechanism of action of SJG-136 both in vitro and in vivo, on the repair of its adducts and mechanism of resistance in cells, and, potentially, on the type of pharmacodynamic assay to be used in clinical trials. SGJ-136 is currently in phase II clinical trials with several groups working on both dimeric cross-linking agents and monoalkylating ligands based on the PBD alkylating moiety. This study suggests subtle differences between the DNA binding of SJG-136 and the C2 unsubstituted analogue DSB-120 that are likely to be the origins of the differences in potency. Confirmation of the stereochemical configuration at the C11 position (particularly in the intrastrand adduct) provides confirmation of binding orientation that was previously only speculation in the HPLC MS study. Together, these observations are likely to be of value in the development of third-generation PBD-based cross-linkers and monoalkylating analogues.

Manipulative Interplay of Two Adozelesin Molecules with d(ATTAAT)2 Achieving Ligand-Stacked Watson–Crick and Hoogsteen Base-Paired Duplex Adducts

Previous structural studies of the cyclopropapyrroloindole (CPI) antitumor antibiotics have shown that these ligands bind covalently edge-on into the minor groove of double-stranded DNA. Reversible covalent modification of the DNA via N3 of adenine occurs in a sequence-specific fashion. Early nuclear magnetic resonance and molecular modeling studies with both mono- and bis-alkylating ligands indicated that the ligands fit tightly within the minor groove, causing little distortion of the helix. In this study, we propose a new binding model for several of the CPI-based analogues, in which the aromatic secondary rings form π-stacked complexes within the minor groove. One of the adducts, formed with adozelesin and the d(ATTAAT)(2) sequence, also demonstrates the ability of these ligands to manipulate the DNA of the binding site, resulting in a Hoogsteen base-paired adduct. Although this type of base pairing has been previously observed with the bisfunctional CPI analogue bizelesin, this is the first time that such an observation has been made with a monoalkylating nondimeric analogue. Together, these results provide a new model for the design of CPI-based antitumor antibiotics, which also has a significant bearing on other structurally related and structurally unrelated minor groove-binding ligands. They indicate the dynamic nature of ligand-DNA interactions, demonstrating both DNA conformational flexibility and the ability of two DNA-bound ligands to interact to form stable covalent modified complexes.

Abstract 3506: Rules of DNA adduct formation for pyrrolobenzodiazepine (PBD) dimers

Abstract Pyrrolo[2,1-c][1,4]benzodiazepine (PBD) dimers are synthetic sequence-selective DNA minor-groove cross-linking agents developed from the tricyclic anthramycin family of naturally-occurring antitumor agents. The “lead” agent, SJG-136, is about to enter Phase II clinical trials as a “first-in-class” sequence-selective DNA minor-groove binding agent. PBDs possess electrophilic imine or carbinolamine moieties that form covalent aminal linkages with guanine C2-NH2 functionalities within the DNA minor groove. Previous studies have shown that PBD dimers with odd numbers of methylenes (e.g., n = 3 or 5) in the -O-(CH2)n-O- linker connecting the two PBD units through their C8/C8′-positions are the most potent interstrand cross-linking agents. We have developed a reversed-phase ion pair HPLC/MS method as a tool to evaluate the interaction of DNA-binding molecules with oligonucleotides of varying length and sequence. Based on this methodology, we recently reported that SJG-136 (n = 3) can form three different types of adducts: inter- and intrastrand cross-linked duplex adducts, and mono-alkylated adducts) (J. Am. Chem. Soc., 2009, 131, 13756-13766). Extending these studies to PBD dimers with longer central linkers (e.g., ELB-21 [n = 5] and DRG-16 [n = 5]), we have found that the type and distribution of the various adducts appear to depend on (1) the positioning of the two reactive guanine bases on the same or opposite strands, (2) their separation (typically the number of base pairs, usually ATs, between them) and (3) the length of the C8/C8′-linker connecting the two PBD units in the dimers. Based on these data, a set of rules have emerged that can be used to predict the binding behaviour of a particular PBD dimer towards a given DNA sequence. For example, for PBD dimers containing an -O-CH2CH2CH2-O- linker (i.e., n = 3) between the C8/C8′-positions the following rules apply: Rule 1 - For binding sequences with two guanines separated by two A/T base pairs (ideal sequence length for “n = 3” dimers according to previous modeling studies), there is a preference for interstrand rather than intrastrand cross-link formation; Rule 2 - For shorter binding sequences (e.g., two guanines separated by a single A/T base pair), interstrand cross-linking is preferred over intrastrand, although the rate of reaction is significantly slower; Rule 3 - For longer sequences (e.g., two guanines separated by three A/T base pairs), intrastrand cross-link formation is preferred and the rate of formation is faster. Similar rules have been found to apply to “n = 5” PBD dimers (e.g., ELB-21 and DRG-16) in relation to sequences with two guanines separated by 2, 3 or 4 base pairs. These results, and the simple rules that appear to explain them, suggest that it may be possible to design PBD dimers to target specific DNA sequences. Note: This abstract was not presented at the AACR 101st Annual Meeting 2010 because the presenter was unable to attend. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 3506.

Origins of the Distortions in the Base Pair Step Adjacent to Platinum Anticancer Drug−DNA Adducts. Fundamental NMR Solution Studies Utilizing Right-Handed Cross-Link Models Having 5′- and 3′-Flanking Residues

For DNA duplexes, the Lippard laboratory has shown that the XG* base pair (bp) step has a very unusual slide and shift, where G* is a G platinated at N7 by di- or monofunctional platinum anticancer drugs. One approach toward understanding the cause of this important unexpected XG* distortion is to examine single-strand (ss) oligonucleotide (oligo) models. Both duplex and ss XG*G* models of the key G*G* cross-link formed by cisplatin have the HH1 conformation with head-to-head bases. Cross-links have R canting (3'-G* H8 atom toward 5'-G*) in duplexes and L canting (5'-G* H8 atom toward 3'-G*) in ss models. However, dynamic motion in solution makes the ss features difficult to define. Thus, we employ less dynamic cross-link models such as (R,S,S,R)-BipPt(d(TG*G*)) and (R,S,S,R)-BipPt(d(pG*G*TTT)), the first examples of an HH1 conformer with R canting for ss oligos longer than d(GpG) (Bip = 2,2'-bipiperidine). In these, the 5'-T residue decreases R canting (indicating steric clashes with the 5'-G*) and the less bulky 5'-phosphate group forms a H-bond to HN-Pt (indicating that R canting allows H-bonding). We conclude that the 5'-X residue in duplex adducts changes its position from that in B form DNA to avoid steric clashes with the 5'-G* and the carrier ligand and secondarily to form a Watson-Crick base pair. These features, possibly aided by weak carrier-ligand H-bonding, lead to the relatively unusual features distinctive to the "Lippard bp step".

Effect of bulky lesions on DNA: solution structure of a DNA duplex containing a cholesterol adduct.

The three-dimensional solution structure of two DNA decamers of sequence d(CCACXGGAAC)-(GTTCCGGTGG) with a modified nucleotide containing a cholesterol derivative (X) in its C1 '(chol)alpha or C1 '(chol)beta diastereoisomer form has been determined by using NMR and restrained molecular dynamics. This DNA derivative is recognized with high efficiency by the UvrB protein, which is part of the bacterial nucleotide excision repair, and the alpha anomer is repaired more efficiently than the beta one. The structures of the two decamers have been determined from accurate distance constraints obtained from a complete relaxation matrix analysis of the NOE intensities and torsion angle constraints derived from J-coupling constants. The structures have been refined with molecular dynamics methods, including explicit solvent and applying the particle mesh Ewald method to properly evaluate the long range electrostatic interactions. These calculations converge to well defined structures whose conformation is intermediate between the A- and B-DNA families as judged by the root mean square deviation but with sugar puckerings and groove shapes corresponding to a distorted B-conformation. Both duplex adducts exhibit intercalation of the cholesterol group from the major groove of the helix and displacement of the guanine base opposite the modified nucleotide. Based on these structures and molecular dynamics calculations, we propose a tentative model for the recognition of damaged DNA substrates by the UvrB protein.

Interaction between ciprofloxacin and DNA mediated by Mg2+-ions

The oligonucleotide duplex d(C1C2T3C4G5C6T7C8T9C10)·d(G11A12G13A14G15C16G17A18G19G20) has been titrated with the fluoroquinolone ciprofloxacin (CFX) using 1H 1D and 2D NMR spectroscopy to monitor the interaction pattern. The assignments of key intermolecular proton–proton crosspeaks in the NOESY map between the fluoroquinolone and the duplex prove the existence of predominantly minor groove CFX–duplex interactions. When MgCl2 was added to a solution of 1:1 CFX–duplex until a final concentration of Mg2+–[PO4]=1, the binding pattern did not change significantly. Theoretical calculations (Docking) carried out on a model of the ternary CFX–Mg2+–duplex adduct using a key interproton distance derived from the NOESY map as anchoring, produced an energetically favourable orientation of the CFX–Mg2+ in the minor groove.

Evidence for a common molecular basis for sequence recognition of N3-guanine and N3-adenine DNA adducts involving the covalent bonding reaction of (+)-CC-1065.

The antitumor antibiotic (+)-CC-1065 can alkylate N3 of guanine in certain sequences. A previous high-field 1H NMR study on the (+)-CC-1065d[GCGCAATTG*CGC]2 adduct (* indicates the drug alkylation site) showed that drug modification on N3 of guanine results in protonation of the cross-strand cytosine [Park, H.-J.; Hurley, L. H. J. Am. Chem. Soc. 1997, 119, 629.]. In this contribution we describe a further analysis of the NMR data sets together with restrained molecular dynamics. This study provides not only a solution structure of the (+)-CC-1065(N3-guanine) DNA duplex adduct but also new insight into the molecular basis for the sequence-specific interaction between (+)-CC-1065 and N3-guanine in the DNA duplex. On the basis of NOESY data, we propose that the narrow minor groove at the 7T8T step and conformational kinks at the junctions of 16C17A and 18A19T are both related to DNA bending in the drugDNA adduct. Analysis of the one-dimensional 1H NMR (in H2O) data and rMD trajectories strongly suggests that hydrogen bonding linkages between the 8-OH group of the (+)-CC-1065 A-subunit and the 9G10C phosphate via a water molecule are present. All the phenomena observed here in the (+)-CC-1065(N3-guanine) adduct at 5'-AATTG* are reminiscent of those obtained from the studies on the (+)-CC-1065(N3-adenine) adduct at 5'-AGTTA*, suggesting that (+)-CC-1065 takes advantage of the conformational flexibility of the 5'-TPu step to entrap the bent structure required for the covalent bonding reaction. This study reveals a common molecular basis for (+)-CC-1065 alkylation at both 5'-TTG* and 5'-TTA*, which involves a trapping out of sequence-dependent DNA conformational flexibility as well as sequence-dependent general acid and general base catalysis by duplex DNA.

Dramatic 5'-residue effect on conformer distribution of short oligonucleotide retro models of the cisplatin-DNA cross-link: implications for the Lippard and cross-link distorted base pair steps present in cisplatin-DNA duplex adducts.

The cisplatin anticancer drug preferentially attacks the GG sequence of DNA duplexes. Virtually all DNAs containing the key G*G* lesion (G* = N7 platinated G) have large distortions in the cross-link (G*G*) base pair (bp) step and also in the adjacent Lippard (XG*) bp step, making the adducts very different from B-form DNA in the XG*G* region. The XG*G* strand in duplexes also differs in several ways from single-strand (ss) models with G*G* and XG*G* sequences. In the duplex, the X residue has an N sugar, the 5'-G* and 3'-G* bases have slight "R" canting (3'-G* H8 atom toward the 5'-G* base), and there is no or weak H-bonding by the NH3 ligands. In most XG*G* ss models, X has an S sugar, the 5'-G* base normally cants strongly toward the 3'-G* base (L canting), and the NH3 forms an H-bond. Well-defined ss models exist in the solid state, but dynamic motion obscures the properties of the ss models in solution. In this work, we employ retro models (better defined, less dynamic ss models) to understand the differences between duplex and ss models. The retro models in this study lack carrier ligand NH's, thus eliminating H-bonding. To correlate previous ss solid-state models with our solution work, we constructed hybrid molecules by overlaying parts of known structures. The combined model and experimental information indicates that the X N-pucker is not favorable in L-canted ss models, that X residue steric effects (not H-bonding) favor L canting in ss models, that X N-pucker is needed for favorable WC hydrogen bonding and stacking interactions in duplexes, and that X N-pucker minimizes X base clashes with bases in the complementary strand in duplexes. The R canting minimizing clashes between the X and G* residues of the Lippard bp step (independent of X pucker) and the repositioning of the X residue base caused by the change from S-pucker to N-pucker together lead to the unusual features of the Lippard bp step in the duplex.

DNA structural perturbation induced by the CPI-derived DNA interstrand cross-linker: molecular mechanisms for the sequence specific recognition

The highly potent cytotoxic DNA-DNA cross-linker consists of two cyclopropa[c]pyrrolo[3, 4-3]indol-4(5H)-ones indoles [(+)-CPI-I] joined by a bisamido pyrrole (abbreviated to "Pyrrole"). The Pyrrole is a synthetic analog of Bizelesin, which is currently in phase II clinical trials due to its excellent in vivo antitumor activity. The Pyrrole has 10 times more potent cytotoxicity than Bizelesin and mostly form DNA-DNA interstrand cross-links through the N3 of adenines spaced 7 bp apart. The Pyrrole requires a centrally positioned GC base pair for high cross-linking reactivity (i.e., 5'-T(A/T)2G(A/T)2A*-3'), while Bizelesin prefers purely AT-rich sequences (e.g., 5'-T(A/T)4 or 5A*-3', where T represents the cross-strand adenine alkylation and A* represents an adenine alkylation) (Park et al., 1996). In this study, the high-field 1H-NMR and rMD studies are conducted on the 11-mer DNA duplex adduct of the Pyrrole where the 5'-TTAGTTA*-3' sequence is cross-linked by the drug. A severe structural perturbation is observed in the intervening sequences of cross-linking site, while a normal B-DNA structure is maintained in the region next to the drug-modified adenines. Based upon these observations, we propose that the interplay between the bisamido pyrrole unit of the drug and central G/C base pair (hydrogen-bonding interactions) is involved in the process of cross-linking reaction, and sequence specificity is the outcome of those interactions. This study suggests a mechanism for the sequence specific cross-linking reaction of the Pyrrole, and provides a further insight to develop new DNA sequence selective and distortive cross-linking agents.

Solution structure of a guanine-N7-linked complex of the mitomycin C metabolite 2,7-diaminomitosene and DNA. Basis of sequence selectivity.

2,7-Diaminomitosene (2,7-DAM), the major metabolite of the antitumor antibiotic mitomycin C, forms DNA adducts in tumor cells. 2,7-DAM was reacted with the deoxyoligonucleotide d(GTGGTATACCAC) under reductive alkylation conditions. The resulting DNA adduct was characterized as d(G-T-G-[M]G-T-A-T-A-C-C-A-C) (5), where [M]G stands for a covalently modified guanine, linked at its N7-position to C10 of the mitosene. The adducted oligonucleotide complements with itself, retaining 2-fold symmetry in the 2:1 drug-duplex complex, and provides well-resolved NMR spectra, amenable for structure determination. Adduction at the N7-position of G4 ([M]G, 4) is characterized by a downfield shift of the G4(H8) proton and separate resonances for G4(NH(2)) protons. We assigned the exchangeable and nonexchangeable proton resonances of the mitosene and the deoxyoligonucleotide in adduct duplex 5 and identified intermolecular proton-proton NOEs necessary for structural characterization. Molecular dynamics computations guided by 126 intramolecular and 48 intermolecular distance restraints were performed to define the solution structure of the 2,7-DAM-DNA complex 5. A total of 12 structures were computed which exhibited pairwise rmsd values in the 0.54-1.42 A range. The 2,7-DAM molecule is anchored in the major groove of DNA by its C10 covalently linked to G4(N7) and is oriented 3' to the adducted guanine. The presence of 2,7-DAM in the major groove does not alter the overall B-DNA helical structure. Alignment in the major groove is a novel feature of the complexation of 2,7-DAM with DNA; other known major groove alkylators such as aflatoxin, possessing aromatic structural elements, form intercalated complexes. Thermal stability properties of the 2,7-DAM-DNA complex 5 were characteristic of nonintercalating guanine-N7 alkylating agents. Marked sequence selectivity of the alkylation by 2,7-DAM was observed, using a series of oligonucleotides incorporating variations of the 5'-TGGN sequence as substrates. The selectivity correlated with the sequence specificity of the negative molecular electrostatic potential of the major groove, suggesting that the alkylation selectivity of 2,7-DAM is determined by sequence-specific variation of the reactivity of the DNA. The unusual, major groove-aligned structure of the adduct 5 may account for the low cytotoxicity of 2,7-DAM.

Chiroptical, ESMS and NMR spectroscopic study of the interaction of enantiopure lanthanide complexes with selected self-complementary dodecamer oligonucleotides†

The interaction of Δ- and Λ-Eu and Yb cationic complexes bearing an N-methylphenanthridinium chromophore with [(CG)6]2, [(AT)6]2 and [CGCGAATTCGCG]2 has been interrogated by ESMS, 1H NMR, absorption, difference circular dichroism, fluorescence quenching and Ln luminescence emission spectroscopy. Stepwise complexation occurs with up to 3 ∶ 1 limiting stoichiometry for [(CG)6]2 and 2 ∶ 1 for [CGCGAATTCGCG]2, as indicated by absorption measurements and direct ESMS observation of the non-covalent duplex adducts. Binding to [(CG)6]2 occurred with an affinity of 8.7 × 106 M−1 duplex−1 for the Δ-Eu isomer which was 50 times greater than to [(AT)6]2. A primary component of the free energy of binding has been linked to an intercalative interaction, supported by absorption, ICD and fluorescence quenching characteristics. The lanthanide coordination environment and local helicity remain unchanged but the oligonucleotide underwent distinctive changes in local helicity and pitch which were sensitive to the handedness of the Ln complex and in certain cases to the nature of the Ln ion (Yb vs. Eu). With [CGCGAATTCGCG]21H NMR TOCSY and NOESY analysis in the presence of increasing concentrations of the Gd analogues revealed that the most avid complex binding site was located in the centre of the oligonucleotide, with the Δ-isomer binding more strongly.

Determination of the structural role of the linking moieties in the DNA binding of adozelesin.

Adozelesin (formerly U73975, The Upjohn Co.) is a monofunctional DNA alkylating analogue of the antitumor antibiotic (+)-CC-1065. Adozelesin consists of a cyclopropa[c]pyrrolo[3,2-e]indol-4(5H)-one (CPI) alkylating subunit of (+)-CC-1065 and a indole and benzofurans subunit replacing the more complex pyrroloindole B and C subunits, respectively, of (+)-CC-1065. Previous studies have shown that adozelesin forms a reversible covalent DNA duplex adduct via a reaction between the N3 of adenine and the cyclopropyl of the cyclopropapyrroloindole (CPI) subunit. Gel electrophoresis studies have shown that adozelesin, like all the monofunctional (CPI)-based antitumor antibiotics, has a sequence preference for 5'-TTA-3' [the asterisk () indicates covalently modified base]. Molecular-modeling studies have shown that the bound adozelesin ligand spans a total of five base pairs including the modified adenine. These studies have also indicated that, owing to the orientation of the ligand within the base minor groove, there should be an overall preference for sequences rich in A.T base pairs, thus avoiding steric crowding around the exocyclic NH(2) of any guanines present. In this study, we have prepared and studied, by high-field NMR and restrained molecular mechanics (rMM) and dynamics (rMD), the duplex adduct formed between adozelesin and 5'-CGTAAGCGCTTACG-3'. Previous molecular-modeling studies suggested that this sequence should be less preferred, since the two GC base pairs should lead to extensive steric crowding within the adduct, and this hypothesis has, however, never been supported by DNA-footprinting data. (1)H NMR of the adozelesin duplex adduct has reveals that, although Watson-Crick base pairing is maintained throughout the DNA duplex, there is significant distortion around the central base pairs. This distortion is the result of strong hydrogen-bonding between the amide linker of the indole and benzofuran subunits, and the carbonyl of a central thymine base and second, weaker, hydrogen bond to the exocyclic NH(2) of the central guanine was also observed. (1)H NMR and rMD also indicate that, to accommodate this hydrogen-bond system, the bound adozelesin is not positioned centrally within the minor groove but pushed toward the modified DNA strand. Previous studies on the dimeric CPI analogue bizelesin have indicated the important role the ureylene linker plays in the DNA binding. This study indicates that a similar situation exists in the reaction of adozelesin with double-stranded DNA and provides a possible explanation into the unpredicted sequence selectivity of these ligands.

Molecular Basis for the DNA Sequence Selectivity of Ecteinascidin 736 and 743: Evidence for the Dominant Role of Direct Readout via Hydrogen Bonding

The marine natural product ecteinascidin 743 (Et 743) is currently in phase II clinical trials. We have undertaken parallel structural and modeling studies of an Et 743−(N2-guanine) 12-mer DNA adduct and an adduct involving the structurally related Et 736 of the same sequence in order to ascertain the structural basis for the ecteinascidin−DNA sequence selectivity. In contrast to the C-subunit differences found in Et 736 and Et 743, they have identical A−B-subunit scaffolds, which are the principal sites of interaction with DNA bases. These identical scaffolds generate parallel networks of drug−DNA hydrogen bonds that associate the drugs with the three base pairs at the recognition site. We propose that these parallel hydrogen bonding networks stabilize the Et 736 and Et 743 A- and B-subunit prealkylation binding complex with the three base pairs and are the major factors governing sequence recognition and reactivity. The possibility that a unique hydrogen-bonding network directs the course of sequence recognition was examined by first characterizing the hydrogen-bonding substituents using 1H NMR properties of the exchangeable protons attached to the hydrogen-bond donor and other protons near the proposed acceptor. Using these experimental findings as indicators of hydrogen bonding, Et 736−12-mer duplex adduct models (binding and covalent forms) containing the favored sequences 5‘-AGC and 5‘-CGG were examined by molecular dynamics (MD) in order to evaluate the stability of the hydrogen bonds in the resulting conformations. The MD-generated models of these favored sequences display optimal donor/acceptor positions for maximizing the number of drug−DNA hydrogen bonds prior to covalent reaction. The results of MD analysis of the carbinolamine (binding) forms of the sequences 5‘-GGG (moderately reactive) and 5‘-AGT (poorly reactive) suggested reasons for their diminished hydrogen-bonding capability. These experimental and modeling results provide the structural basis for the following sequence specificity rules: For the target sequence 5‘-XGY, the favored base to the 3‘-side, Y, is either G or C. When Y is G, then a pyrimidine base (T or C) is favored for X. When Y is C, a purine (A or G) is favored for X.

Solution structure of the aminofluorene [AF]-intercalated conformer of the syn-[AF]-C8-dG adduct opposite dC in a DNA duplex.

We report below on a conformational equilibrium between AF-intercalated and AF-external states in slow exchange for the [AF]dG lesion positioned opposite dC in the d(C-[AF]G-C).d(G-C-G) sequence context. The slow exchange between states is attributed to interconversion between syn glycosidic torsion angle in the AF-intercalated and anti torsion angle in AF-external conformers of the [AF]dG opposite dC containing duplex. The present paper describes an NMR-molecular mechanics study that defines the solution structure of the AF-intercalated conformer for the case of [AF]dG adduct positioned opposite dC in the d(C-[AF]G-C).d(G-C-G) sequence context. The structure is of the base displacement-intercalation type where the aminofluorene ring is intercalated into the helix between intact Watson-Crick dG.dC base pairs, which results in a displacement of the modified guanine ring into the major groove where it stacks with the major groove edge of its 5'-flanking cytosine in the adduct duplex. The conformational equilibrium between AF-intercalated conformer (approximately 70%) with a syn alignment and AF-external conformer (approximately 30%) with an anti alignment for the [AF]dG adduct positioned opposite dC in the d(C-[AF]G-C).d(G-C-G) sequence context can be contrasted with our earlier demonstration that the population is 100% for the AP-intercalated conformer with a synalignment at the N-(deoxyguanosin-8-yl)-2-aminopyrene ([AP]dG) adduct site positioned opposite dC in the same sequence context [Mao, B., Vyas, R. R., Hingerty, B. E., Broyde, S., Basu, A. K., and Patel, D. J. (1996) Biochemistry, 35, 12659-12670]. This shift in population may reflect the much larger size of the pyrenyl ring of the [AP]dG adduct compared to the fluorenyl ring of the [AF]dG adduct which in turn might provide for a greater overlap of the aromatic amine with the flanking base pairs in the intercalated conformer of the former adduct in DNA.

Unrestrained Molecular Dynamics of Photodamaged DNA in Aqueous Solution

The structural impact of photolesions on DNA was investigated by carrying out molecular dynamics simulations of the T5/T6 cis-syn dimer (A), T5/T6 6-4 adduct (B), and native decamer (C) of d(CGCATTACGC)2. For the cis-syn dimer containing decamer the simulations provided a structure very similar to that derived with experimental NOE data. Specifically, distortions of the simulated cis-syn dimer duplex (A) were primarily localized at the lesion site and had an overall bend similar to that of the refined NMR structure. For the simulated 6-4 adduct duplex (B), although there was considerable coincidence with the NMR derived structure, an intact hydrogen bond/dipole−dipole interaction between the T6 pyrimidone carbonyl oxygen and A15/NH2 was observed, which was not found in the NMR derived structure. This hydrogen bond leads to a structure that seems to better account for some of the unusual experimental NOEs than the refined NMR structure. In addition, for the 6-4 adduct (B), the overall bend was significantly smaller than that suggested by the NMR derived structure. This difference in overall bend is presumably a result of differences in the torsion angles at the lesion site for the simulated vs the experimental structure. Overall these simulations agreed well with the features of the spectroscopically determined structures and reveal the power of molecular dynamics methods with explicit solvation and accurate representation of long-range electrostatic interactions to usefully model noncanonical forms of DNA.

Differential Hydration Thermodynamics of Stereoisomeric DNA−Benzo[a]pyrene Adducts Derived from Diol Epoxide Enantiomers with Different Tumorigenic Potentials

A combination of UV spectroscopy, calorimetry, and density techniques were used to characterize the thermodynamics of complexes with covalently bound hydrophobic pyrenyl residues in the minor groove of DNA undecamer duplexes. The control duplex d(CCATCG*CTACC)/d(GGTAGCGATGG) and two adduct duplexes in which the chiral (+)-anti-BPDE and (−)-anti-BPDE (the 7R,8S,9S,10R- and 7S,8R,9R,10S-enantiomers of 7r,8t-dihydroxy-9t,10t-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene) had been reacted covalently with the exocyclic amino group of the guanine residue G* were studied (designated as the (+)- and (−)-BPDE duplexes, respectively). Both of the BPDE-modified DNA duplexes exhibit lower helix-coil transition temperatures than the control duplex. The complete thermodynamic profiles (ΔV, ΔH, ΔG, ΔS, and ΔnNa+) for the formation of each duplex were determined at 20 °C. Duplex formation is primarily enthalpy driven, and is accompanied by an uptake of both counterions and water molecules (negative ΔV). Relative to the unmodified duplex, the differential thermodynamic profiles of each covalent adduct duplex reveal an enthalpy−entropy compensation; the ΔΔV value is only marginally smaller for the (−)-BPDE−DNA than for the unmodified duplex, but the uptake of water is nearly 50% greater for the (+)-BPDE duplex. Correlation of the thermodynamic data with the known NMR solution conformations of the BPDE−DNA complexes (de los Santos et al. Biochemistry 1992, 31, 5245) suggests that these differential thermodynamic parameters, together with the similar values for the uptake of counterions, correspond to a differential hydration of the BPDE residues that are exposed to solvent while in the minor groove of B-DNA. The formation of the (+)-BPDE duplex results in a greater immobilization of structural water than in the case of the (−)-BPDE duplex; these results suggest that the bent conformation at the lesion site apparently gives rise to an enhanced exposure of the hydrophobic polycyclic aromatic moiety of the covalently bound BPDE residue to the aqueous solvent.

Solution conformation of the N-(deoxyguanosin-8-yl)-1-aminopyrene ([AP]dG) adduct opposite dC in a DNA duplex.

Combined NMR-molecular mechanics computational studies were undertaken on the C8-deoxyguanosine adduct formed by the carcinogen 1-nitropyrene embedded in the d(C5-[AP]G6-C7).d(G16-C17-G18) sequence context in a 11-mer duplex, with dC opposite the modified deoxyguanosine. The exchangeable and nonexchangeable protons of the aminopyrene moiety and the nucleic acid were assigned following analysis of two-dimensional NMR data sets in H2O and D2O solution. There was a general broadening of several proton resonances for the three nucleotide d(G16-C17-G18) segment positioned opposite the [AP]dG6 lesion site resulting in weaker NOEs involving these protons in the adduct duplex. The solution conformation of the [AP]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 spectra as restraints in molecular mechanics computations in torsion angle space. The aminopyrene ring of [AP]dG6 is intercalated into the DNA helix between intact Watson-Crick dC5.dG18 and dC7.dG16 base pairs. The modified deoxyguanosine ring of [AP]dG6 is displaced into the major groove and stacks with the major groove edge of dC5 in the adduct duplex. Both carbon and proton chemical shift data for the sugar resonances of the modified deoxyguanosine residue are consistent with a syn glycosidic torsion angle for the [AP]dG6 residue. The dC17 base on the partner strand is displaced from the center of the helix toward the major groove as a consequence of the aminopyrene ring intercalation into the helix. This base-displaced intercalative structure of the [AP]dG.dC 11-mer duplex exhibits several unusually shifted proton resonances which can be accounted for by the ring current contributions of the deoxyguanosinyl and pyrenyl rings of the [AP]dG6 adduct. In summary, intercalation of the aminopyrene moiety is accompanied by displacement of both [AP]dG6 and the partner dC17 into the major groove in the [AP]dG.dC 11-mer duplex.

Solution conformation of the (-)-cis-anti-benzo[a]pyrenyl-dG adduct opposite dC in a DNA duplex: intercalation of the covalently attached BP ring into the helix with base displacement of the modified deoxyguanosine into the major groove.

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.

Determination of the Structural Role of the Internal Guanine-Cytosine Base Pair in Recognition of a Seven-Base-Pair Sequence Cross-Linked by Bizelesin

Bizelesin (formerly U77,779, The Upjohn Co.) is a bifunctional DNA cross-linking antitumor antibiotic consisting of two open-ring homologs of the (+)-CC-1065 cyclopropa[c]pyrrolo[3,2-e]indol-4(5H)-one (CPI) subunits connected by a rigid linking moiety. Previous studies have shown that Bizelesin most often forms an interstrand cross-link through the N3 of two adenines 6 base pairs (bp) apart (inclusive of the modified adenines). However, gel electrophoresis studies have also indicated that Bizelesin forms 7-bp cross-links in specific sequences. In most of these sequences the cross-linked adenines represent the only possible cross-link site (i.e., no 6-bp site is available); however, in several sequences, a 7-bp sequence is selected in overwhelming preference to a possible 6-bp sequence. In this study, we demonstrate the unique requirement for a G.C base pair within this sequence and the critical presence of the exocyclic 2-amino group of guanine. In a subsequent two-dimensional 1H-NMR study that concentrates on the 7-bp cross-link formed with the sequence 5'-TTAGTTA-3', the role of the central G.C base pairs in the formation of a 7-bp cross-link is probed. 1H-NMR analysis coupled with restrained molecular dynamics (rMD) provides evidence for distortion around the covalently modified adenines. Because of this distortion, the modified bases are twisted toward the center of the duplex adduct, effectively reducing the cross-linked distance. The rMD study also indicates that a hydrogen bond is formed between the exocyclic amine of the central guanine and the carbonyl of the ureylene linker. On the basis of the observation of the distortion in the duplex and the hydrogen bonding between the drug and DNA, it is possible to speculate on the role of the central G.C bases in this sequence preference and propose a mechanism by which Bizelesin forms a 7-bp rather than a 6-bp cross-link with this sequence.