Molecular recognition and binding of a GC site-avoiding thiazole-lexitropsin to the decadeoxyribonucleotide d-[CGCAATTGCG]2: 1H-NMR evidence for thiazole intercalation

Sequence Selective Molecular Recognition of Long DNA Sequences by Oligomethylene-Linked Oligoimidazole Analogs of Distamycin

We have studied a series of homologous N-to-N oligomethylene linked bis(diimidazole) analogs 4a-f (number of methylene groups = 1 to 6, re spectively) and a dipicolinamide congener 4g that bind to long GC-containing sequences of DNA. Results from an ethidium binding assay reveal that, for 4a-f, the compounds with an even number of methylene groups have larger ap parent binding constants, Kapp, than those with an odd number. The Kapp values of the compounds with an odd number of methylene groups are close to that of their monomeric analog 3 suggesting that they may be binding to DNA in a monodentate fashion. The binding of these compounds to T4 DNA and their larger binding constants for poly(dG-dC) over poly(dA-dT) indicated minor groove binding selectivity and tolerance for GC sequences. The ability of com pounds 4b-f to bind selectively to DNA was illustrated by an MPE-Fe(II) foot printing study which showed that compound 4f gave the most distinct foot prints. CD titration studies on compounds 4b, d, and f further demonstrated the GC tolerance of these compounds and that they can bind to 7 ~ 8 base pairs of DNA in a bidentate fashion. The minor groove and bidentate bind ing of the ethylene linked compound 4b on the underlined sequence of

Intercalation, DNA kinking, and the control of transcription

Biological processes involved in the control and regulation of transcription are dependent on protein-induced distortions in DNA structure that enhance the recruitment of proteins to their specific DNA targets. This function is often accomplished by accessory factors that bind sequence specifically and locally bend or kink the DNA. The recent determination of the three-dimensional structures of several protein-DNA complexes, involving proteins that perform such architectural tasks, brings to light a common theme of side chain intercalation as a mechanism capable of driving the deformation of the DNA helix. The protein scaffolds orienting the intercalating side chain (or side chains) are structurally diverse, presently comprising four distinct topologies that can accomplish the same task. The intercalating side chain (or side chains), however, is exclusively hydrophobic. Intercalation can either kink or bend the DNA, unstacking one or more adjacent base pairs and locally unwinding the DNA over as much as a full turn of helix. Despite these distortions, the return to B-DNA helical parameters generally occurs within the adjacent half-turns of DNA.

DNA recognition by lexitropsins, minor groove binding agents

Consideration is given to alternative approaches to the development of DNA sequence selective binding agents because of their potential applications in diagnosis and treatment of cancer as well as in molecular biology. The concept of lexitropsins, or information-reading molecules, is introduced within the antigene strategy as an alternative to, and complementary with, the antisense approach for cellular intervention and gene control. The chemical, physical and pharmacological factors involved in the design of effective lexitropsins are discussed and illustrated with experimental results. Among the factors contributing to the molecular recognition processes are: the presence and disposition of hydrogen bond accepting and donating groups, ligand shape, chirality, stereochemistry, flexibility and charge. For longer ligands, such as are required to target unique sequences in biological systems (14-16 base pairs), the critical feature is the phasing or spatial correspondence between repeat units in the ligand and the receptor. The recently discovered 2:1 lexitropsin-DNA binding motif provides a further refinement in molecular recognition in permitting discrimination between GC and CG base pairs. The application of these factors in the design and synthesis of novel agents which exhibit anticancer, antiviral and antiretroviral properties, and inhibition of critical cellular enzymes including topoisomerases is discussed. The emerging evidence of a relationship between sequence selectivity of the new agents and the biological responses they invoked is also described.

Quantum chemical and molecular mechanics studies on the binding of stereoisomers of the oligopeptide antibiotics amidinomycin and noformycin to the minor groove of B-DNA

Ab initio calculations (Hartree-Fock) using the 6-31G basis set have been performed on two chiral oligopeptide antitumor antibiotics amidinomycin 5 and noformycin 6. The latter are DNA minor groove binding agents related to the A.T recognizing netropsin 4 and distamycin 3 but, unlike the latter, bear stereocenters (two for 5 and one for 6) that may be expected to affect binding to the B-DNA receptor. Geometry optimized conformations, energies and distribution of electrostatic charges within the molecules were derived. The rotational barrier for bond C3-C6 in 6 was calculated to be ca. 6 kcal.mole-1 and the dipole moment for 6 was 7.69D and for 5 was 5.58D. The ab initio derived parameters of the geometry optimized conformations of the different possible stereoisomeric forms of 5 and 6 were used to interpret their different interactions with the minor groove of DNA at both A.T and G.C sequences and the results were compared with molecular mechanics calculations. The order of binding of the four stereoisomers of 5 at the preferred (A.T)n sequences by both ab initio and molecular mechanics calculations is 1S,3R > RR > RS > SS. The predicted energy differences for complexation with DNA of the other stereoisomers from that of 1S,3R are: RR (4.2%); RS (6.7%) and SS (21.5%). In the case of noformycin the 4R structure binds more effectively than the enantiomer. Considerations of phasing in the computed distances between hydrogen bond donating sites in the DNA-bound antibiotics provide further insight into the binding processes. In the complexes of noformycin 6 the N-N1-N4 and N1-N5 distances (9.05 and 9.15 A respectively for 4R-6 and 9.23 and 9.26 A respectively for 4S-6) are close to the optimum value of 9.1 A for effective binding. In the case of amidinomycin 5 the best agreement with the optimum value occurs with the strongest binding diastereomer 1S,3R (N1-N3 = 8.91, N1-N4 = 9.41 A). The unexpected result, consistent in both ab initio and molecular mechanics treatments, is that, in contrast to the cases of kikumycin 1 and anthelvencin 2, the natural 3S configuration of 5 and 4S of 6 do not confer maximal binding efficiency. This suggests that biogenetic factors in the generation of the oligopeptide antibiotics lead to maximum DNA binding in the cases of kikumycin and anthelvencin but not in the cases of amidinomycin and noformycin.

FTIR study of specific binding interactions between DNA minor groove binding ligands and polynucleotides

The use of FTIR spectroscopy is made to study the interactions between polynucleotides and two series of minor groove binding compounds. The latter were developed and described previously as part of an ongoing program of rational design of modified ligands based on naturally occurring pyrrole amidine antibiotic netropsin, and varying the structure of bisbenzimidazole chromosomal stain Hoechst 33258. Characteristic IR absorptions due to the vibrations of thymidine and cytosine keto groups in polynucleotides containing AT and GC base pairs respectively are used to monitor their interaction with the added ligands. Although the two thiazole based lexitropsins based on netropsin structure differ in the relative orientation of nitrogen and sulfur atoms with respect to the concave edge of the molecules, they interact exclusively with the thymidine C2 = O carbonyl groups in the minor groove of the alternating AT polymer as evidenced by specific changes in the IR spectra. In the second series of compounds based on Hoechst 33258, the structure obtained by replacing the two benzimidazoles in the parent compound by a combination of pyridoimidazole and benzoxazole, exhibits changes in the carbonyl frequency region of poly dG.poly dC which is attributed to the ligand interaction at the minor groove of GC base pairs. In contrast, Hoechst 33258 itself interacts only with poly dA.poly dT. Weak or no interaction exists between the ligands and any of the polynucleotides at the levels of the phosphate groups or the deoxyribose units.

Structure and dynamics of ligand-template interactions of topoisomerase inhibitory analogs of Hoechst 33258: high field 1H-NMR and restrained molecular mechanics studies

The binding characteristics of Hoechst 33258 (1), a synthetic bis-benzimidazole, and its structural analog 2, with one of the benzimidazoles replaced by a pyridoimidazole, to the self-complementary decadeoxyribonucleotide sequences d(CGCAATTGCG)2 (A) and d-(CATGGCCATG)2 (B) respectively, were examined using high field 1H-NMR techniques. Selective complexation induced chemical shift changes, the presence of exchange signals and intermolecular NOE contacts between the ligands and the minor groove protons of the oligonucleotides suggest the preferred binding sites as the centrally located AATT segment for complex A1, and the CCAT segment for complex B2. The B-type conformations of the two DNA duplexes are preserved upon complexation, as confirmed by the 2D-NOESY based sequential connectivities involving DNA base and sugar protons. Close intermolecular NOE based contacts between the ligands and their respective DNA sequences were further refined to model the ligand-DNA complexes starting from the computer generated B-type structures for the oligonucleotides. Force field calculations of ligand-DNA interaction energies indicate a more favorable contribution from the van der Waals energy component in the case of complex A1 consistent with its stronger net binding compared with the complex B2. Overall, the incorporation of a pyridinic nitrogen in Hoechst 33258 structure alters its selectivity for base pair recognition from A.T to G.C, resulting largely from the formation of a hydrogen bond between the new basic center and the 2-NH2 group of a guanosine moiety. The rates for the exchange of ligands between the two equivalent binding sites (AATT for 1, and CCAT for 2) of the self-complementary DNA sequences, are estimated from analyses of coalescence of NMR signals to be 189s-1 at 301 K for A1 and 79s-1 at 297 K for B2; which correspond to delta G++ of 13.8 and 18.6 kcal.mol-1 respectively.

DNA-sequence specific recognition by a thiazole analogue of netropsin: a comparative footprinting study

Four different footprinting techniques have been used to probe the DNA sequence selectivity of Thia-Net, a bis-cationic analogue of the minor groove binder netropsin in which the N-methylpyrrole moieties are replaced by thiazole groups. In Thia-Net the ring nitrogen atoms are directed into the minor groove where they could accept hydrogen bonds from the exocyclic 2-amino group of guanine. Three nucleases (DNAase I, DNAase II, and micrococcal nuclease) were employed to detect binding sites on the 160bp tyr T fragment obtained from plasmid pKM delta-98, and further experiments were performed with 117mer and 253mer fragments cut out of the plasmid pBS. MPE.Fe(II) was used to footprint binding sites on an EcoRI/HindIII fragment from pBR322. Thia-Net binds to sites in the minor groove containing 4 or 5 base pairs which are predominantly composed of alternating A and T residues, but with significant acceptance of intrusive GC base pairs. Unlike the parent antibiotic netropsin, Thia-Net discriminates against homooligomeric runs of A and T. The evident preference of Thia-Net for AT-rich sites, despite its containing thiazole nitrogens capable of accepting GC sites by hydrogen bonding, supports the view that the biscationic nature of the ligand imposes a bias due to the electrostatic potential differences in the receptor which favour the ligand reading alternating AT sequences.

Structural and dynamic aspects of non-intercalative (1:1) binding of a thiazole-lexitropsin to the decadeoxyribonucleotide d-[CGCAATTGCG]2: An 1H-NMR and molecular modeling study

The location, orientation and dynamics of a thiazole-containing analogue of distamycin 1 bound to the decadeoxyribonucleotide d-[CGCAATTGCG]2 have been studied by non-exchangable and imino proton NMR resonances of the 1:1 complex. Using NOE difference, COSY and NOESY experiments, lexitropsin (1) was located in the minor groove of DNA at 5'-CAAT sequence. This was concluded by an intermolecular NOE between the ligand and a minor groove A4H2 proton. The NOE cross-correlations in the NOESY map confirmed that the DNA decamer duplex in the 1:1 complex remains in a right-handed B-conformation similar to that in the free decamer. Experiments on non-exchangeable and exchangeable proton NMR resonances placed the N-formylamino terminus of drug 1 on the 5'-C3 nucleotide, while the rest of the molecule extends onto the 5'-AAT sequence. The structural evidence for sequence preferential binding at 5'CAAT rather than 5'AATT suggests this reflects an attempt on the part of the sterically demanding inward directed sulfur of the thiazole to minimize compression by moving part of the molecule to the somewhat wider CG base site. The lack of evidence for a 2:1 drug:DNA complex, in contrast to distamycin, is in accord with this interpretation. The lexitropsin 1 was found to be in an exchange between the equivalent 5'-CAAT sites at a rate of approximately 35S-1 with a delta G degree of 65 +/- 5 kJ mol-1 at 303 K. The experimental data suggests a slide-swing mechanism for this exchange process.