Recognition elements that determine affinity and sequence-specific binding to DNA of 2QN, a biosynthetic bis-quinoline analogue of echinomycin.

Abstract

Footprinting experiments with DNase I provide a starting-point for investigating the molecular basis of nucleotide sequence recognition by 2QN, a bis-quinoline derivative of the quinoxaline antibiotic echinomycin produced by directed biosynthesis in Streptomyces echinatus. Using tyrT DNA molecules variously substituted with inosine and/or 2,6-diaminopurine residues it is shown that the location of the 2-amino group of purine nucleotides in the minor groove of the double helix exerts a dominant influence in determining where the antibiotic will bind, as it does for echinomycin. However, newly created binding sites in DNA molecules substituted with diaminopurine (D), all located round TpD steps, bind 2QN with so much higher affinity than the canonical CpG steps that the latter fail completely to appear as footprints in D-substituted DNA; indeed CpG sequences appear in regions of enhanced susceptibility to nuclease cleavage as do CpI steps in doubly D + I-substituted DNA. Quantitative footprinting plots confirm that sequences surrounding TpD steps bind 2QN several hundred-fold more tightly than do CpG-containing sequences, with dissociation constants of the order of 25 nM. To test the hypothesis that differences in stacking interactions between the chromophores of the drug and the DNA base pairs could account for the differences in binding affinities, models of 2QN bound to two DNA hexamers containing either a central CpG or a central TpD step were built. Calculation of the molecular electrostatic potential (MEP) of 2QN in solution using a continuum method revealed a distinctive pattern that is considered relevant to DNA binding. When the MEPs calculated for the two DNA hexamers in the complexed state were compared, substantial differences were found in the major groove and in the space between the base pairs that is occupied by the chromophores of the drug upon binding. The modelling data support the notion that electrostatic stacking interactions underlie the considerably preferred binding of echinomycin and 2QN around TpD steps rather than CpG steps.