Search

Bis-pyridium oximes and methoximes from a newly synthesized series are weak DNA binders (K = 3.10(4) M-1 under physiological conditions). From the number of binding sites per phosphate, 0.25, the ionic strength dependence of the binding constant and the negative electric dichroism, it is concluded that monointercalation is the mode of association. In contrast to methoxy compounds, the oxime derivatives are able both to induce the mutated "petite" phenotype in yeast S. cerevisiae and to cause "in vitro" extensive condensation of single stranded DNA. This reaction is postulated to be relevant to the mutational process that leads to "peptide" cells. The absence of nuclear mutation is interpreted in terms of sequestration of the drug in mitochondria under the effect of the organelle inner membrane electrochemical potential.

Viscosimetric and kinetic results allow one to characterize three modes of DNA binding in the ellipticine series: (1) Ellipticine and its 9 methoxy derivative, which present maximal DNA lengthening properties and bind DNA through a single step mechanism, can be considered as pure intercalators. (2) Ellipticinium derivatives and short-chain substituted oxazolopyridocarbazoles, which present intermediate DNA lengthening properties, bind DNA through a two-step mechanism, one being intercalation. (3) Long-chain substituted oxazolopyridocarbazole derivatives, which display the smallest DNA lengthening properties, bind DNA through a single-step mechanism, probably resulting from an outside binding mode. The viscosimetric and kinetic results are compared with the thermodynamic results obtained from the temperature dependence of the binding constants. It appears that drugs binding on the outside of the DNA double helix tend to have large enthalpy and small entropy contributions, whereas pure intercalating drugs have contributions from both enthalpy and entropy, with entropy dominating by about 2:1. Drugs showing two binding modes exhibit a continuum between the aforementioned extremes, with no breaks in behavior. From this comparison, a correlation between thermodynamic data and DNA binding modes is proposed. Possible molecular implications of both enthalpy and entropy to DNA binding free energy are discussed.

The temperature-jump (T-jump) method has been used to investigate the binding mechanism to calf-thymus DNA of ellipticine and some of its derivatives. The results show that the plant alkaloid, ellipticine, interacts with DNA at a unique intercalation site whereas most of its synthetic derivatives, such as ellipticinium, 9-hydroxy-ellipticinium and related alkyl-oxazolopyridocarbazoles recognize two distinct DNA sites. Parallel analysis of kinetic data and DNA lengthening abilities of these derivatives suggests that only one of these two DNA sites is an intercalation site. Owing to the determination of the genuine number of drug-DNA complexes (inferred from T-jump experiments) and with the results of thermodynamic investigations (Van't Hoff plots), further characterization of the molecular interactions involved in the binding process was proposed. Thus, the formation of the unique intercalation complex of ellipticine was found to be entropy driven whereas binding of drugs which recognize the second class of binding sites was essentially enthalpy driven. These different thermodynamic behaviors suggest that intercalation essentially results from hydrophobic solvent structure effects in contrast to the second binding mode which principally arises from hydrogen bonding interactions through DNA grooves.

We have determined the acid-base properties of the alkaloid ellipticine, bound to DNA and to micelles and liposomes, taken as models for membranes, in the prospect of characterizing the actual structure of the bound ligand, this being relevant to the mode of action of the drug. The acid-base properties of ellipticine bound to sonicated calf thymus DNA and SDS micelles are similar as regards their pK values and their dependence on NaCl concentrations. This observation is satisfactorily understood in terms of sodium ion condensation around the negative phosphate and sulphate groups. The slope of pK vs log(Na+) is -1, a value predicted by Friedman theory. The pK of ellipticine bound to cationic (CTAB, DDTAB) or to neutral (Triton X100) micelles and to neutral liposomes (PC) is significantly lower than water (7.4), and, in contrast to the situation in DNA and SDS micelles, does not vary with addition of NaCl. Interestingly, this result is good evidence for ellipticine having a specific pK when bound to a hydrophobic structure. This view is likely to hold for ellipticine bound to DNA.

The acid-basic properties of ellipticine have been re-estimated. The apparent pK of protonation at 3 microM drug concentration is 7.4 +/- 0.1. The ellipticine free base (at pH 9, I = 25 mM) intercalates into calf-thymus DNA with an affinity constant of 3.3 +/- 0.2 X 10(5) M-1, and a number of binding sites per phosphate of 0.23. The ellipticinium cation (pH 5, I = 25 mM) binds also to DNA with a constant of 8.3 +/- 0.2 x 10(5) M-1 and at a number of binding sites (n = 0.19). It is postulated that the binding of the drug to DNA at pH 9 is driven by hydrophobic and/or dipolar effects. Even at pH 5, where ellipticine exists as a cation, it is thought that the hydrophobic interaction is the main contribution to binding. The neutral and cationic forms share common binding within DNA sites but yield to structurally different complexes. The free base has 0.04 additional specific binding sites per phosphate. As determined from temperature-jump experiments, the second-order rate constant of the binding of the free base (pH 9) is 3.4 x 10(7) M-1 s-1 and the residence time of the base within the DNA is 8 ms. The rate constant for the binding of the ellipticinium cation is 9.8 x 10(7) M-1 s-1 when it is assumed that drug attachment occurs via a pathway in which the formation of an intermediate ionic complex is not involved (competitive pathway).

Ditercalinium, a 7H-pyridocarbazole dimer designed to bisintercalate into DNA, forms tight ion pairs in water with inorganic and organic anions. The thermodynamics and kinetics of the acetate-ditercalinium pairing has been investigated by means of T-jump spectroscopy. The formation of the pair has a constant estimated to 1000 M-1 and proceeds via a fast two-step mechanism with a relaxation time of 12 microseconds (acetate pH 5) to 50 microseconds (cacodylate, pH 7.5) involving an intermediate solvent-separated ion pair. A strong association of ditercalinium to cardiolipid has been observed and is expected to be involved in the respiratory chain inhibition induced by ditercalinium (unpublished results). Direct estimates of the binding constants of the drug to calf thymus DNA were obtained by means of UV titrations at high drug/base ratio (greater than 0.17). The maximum number of binding sites per base both at pH 5 and pH 7.5 was found to be 0.22, a value consistent with monointercalation as expected from the prediction of Shafer's model for the interaction of bifunctional ligands to DNA. This work also supports the hypothesis that significant ionic binding may account for the ditercalinium/DNA interaction at high base/drug ratios (0.2).

Among the DNA-intercalating drugs in the ellipticinium series, 9-hydroxy derivatives elicit the highest antitumoral properties. In water these drugs display a very low fluorescence quantum yield. Replacement of H2O by D2O partially restores the fluorescence of the ellipticinium chromophore. The possibility that such a proton-exchange mechanism could be involved in a base-recognizing process at the DNA level (and therefore be responsible for some base preference) was examined by direct fluorescence titration in deuterated buffer and DNA/drug fluorescence energy transfer. These experimental approaches provide mutually consistent results showing that the 9-hydroxylated drug recognizes specific DNA sites that are not recognized by the non-hydroxylated drug. When compared to 2-N-methyl ellipticinium, the 2-N-methyl 9-hydroxyellipticinium presents: (1) higher binding constants for each DNA studied; (2) a base dependence of the fluorescence properties of the bound form (fluorescence increment upon DNA binding varying over 5-11); (3) a base dependence of its DNA affinity constants (1.1-3.3 x 10(6) M-1) and of its site size (exclusion parameters varying over 3.0-4.4); (4) a base dependence of its energy transfer from DNA bases. Analysis of the binding data suggests that the 9-hydroxyl group of 2-N-methyl ellipticinium is responsible for a G.C base-pair preference, the preferred binding site being a doublet sequence of two adjacent G.C which could be flanked either by a additional G.C base pair or by an A.T base pair.

Codon repeats encountered in DNA sequences may formally lead to several double-stranded structures of similar stabilities. This is observed in the highly repetitive sequences of some Plasmodium antigens (S-antigen, CS proteins). It is postulated that gene recombination may occur via various heteroduplex molecules thus leading to antigenic polymorphism of Plasmodium parasites.

The temperature-jump method has been used to compare the binding of 2-N methyl ellipticinium (NME) and 2-N methyl 9 hydroxy ellipticinium (NMHE) to three natural DNA's of different AT/GC composition. The relaxation signals, analyzed by the Padé-Laplace method, are characterized by two distinct relaxation times, tau 1 and tau 2, respectively in the 1-4 ms and 20-80 ms range. In the case of the NMHE/DNA interaction, the slower relaxation time tau 2 depends on the DNA composition, as follows: tau 2 (Micrococcus lysodeikticus) greater than tau 2 (Calf thymus) greater than tau 2 (Clostridium perfringens). Contrary to NMHE, NME which does not possess an OH group at the C-9 position, shows no relaxation time dependence upon DNA base composition. The observation of two relaxation times indicates that the binding equilibria are associated with at least two distinct drug/DNA complexes (probably arising from two distinct DNA binding sites). Three kinetic models, involving the formation of a weak intermediate ionic complex, are given to explain the binding reaction between these cationic drugs and the DNA. They allow the determination of the four rate constants associated with the two binding steps and lead to equilibrium association constants in agreement with those obtained from spectroscopic studies. The validity of the models is discussed and it is shown that the best kinetic scheme, for either NMHE or NME, could be that in which the ionic step is not a prerequiste to intercalation. The kinetic results show that the residence time of 9 hydroxy ellipticinium is markedly increased in GC rich DNA's and this could be related to the higher in vitro and in vivo cytotoxic properties of 9 hydroxy substituted ellipticines.