No abstract

By means of titration viscometry a number of distinct modes could be resolved for the interaction between the antibiotic netropsin and DNA species of 50, 58, and 69 mole + (A+T) below r = 0.04 netropsin molecules bound per DNA phosphate group. The number of corresponding binding sites increases with a high power of the (A+T) content. The apparent association constants are very high (greater than 10(6) M-1, some perhaps greater than 10(6) M-1) and also rather different for most of the binding sites. It is suggested that some of these interaction modes differ in the number of hydrogen bonds formed between donors of the ligand and acceptors of the binding sites. The interaction modes were characterized quantitatively by their (species-independent) changes of DNA contour length and by the percentage of local DNA stiffening.

Binding to DNA's of the non-intercalative ligands SN-6999 and SN-18071 has been studied by means of circular dichroism, UV absorption, thermal melting and for SN-6999 by viscosity measurements. Both antitumour drugs show a preference for dA.dT rich DNA's, but the base pair selectivity of SN-18071 is lower as indicated by some affinity to dG.dC containing duplex DNA. The dA.dT base pair specificity of SN-6999 is comparable to that of netropsin. It forms very stable complexes with dA.dT containing duplex DNA and competes with netropsin binding on DNA. The ligands SN-18071 and pentamidine are totally released from their complexes with poly(dA-dT).poly(dA-dT) by competitive netropsin binding. The results demonstrate that hydrogen bonding capacity of the ligand in addition to other factors strongly contribute to the base sequence specificity in the recognition process of the ligand with DNA. A binding model of SN-6999 with five dA.dT pairs in the minor groove of B-DNA is suggested.

Upon interaction of the three anthracycline antibiotics daunomycin, adriamycin, and aclacinomycin A with calf thymus DNA the relative changes of both DNA contour length, delta L/Lo, and persistence length, delta a/ao, have been determined as a function of r, the ratio of bound ligand molecules per DNA mononucleotide. From the r dependence of delta a/ao a measure for the stiffening effect and also the angle gamma of ligand-induced DNA bending could be derived. Experimental basis are titration viscometric measurements upon both low and high molecular weight DNA. It was found that the DNA contour length increases linearly with r by approximately 0.34 nm per bound drug molecule. The comparatively very high DNA stiffening effect measured in solution is understandable as a result of helix clamping by at least two anthracycline groups of sufficient long distance. The variation of gamma on DNA interaction with different anthracycline derivatives find their explanation in terms of different values of the mismatch to in-register binding prior to complex formation. From an analogous interpretation of viscosity measurements by Arcamone and coworkers upon high molecular weight DNA with many anthracycline derivatives it can be concluded that DNA interaction by both amino sugar and 9-acetyl group are responsible for the generation of strong anthracycline binding mediated DNA stiffening effects in solution. (A combined analysis of the viscosity measurements by Cohen & Eisenberg and Armstrong et al. upon DNA interaction with proflavine indicates a very small DNA stiffening effect, gamma = 6.7 sigma and a helix elongation by 0.35 nm per bound ligand molecule.)

No abstract