Revisiting the salt dependence of the drug-DNA association process: A Poisson-Boltzmann analysis

Wednesday, September 12, 2007 - 11:15am - 12:15pm
Lind 409
Marcia Fenley (Florida State University)
The proper modeling of salt-mediated electrostatic interactions is essential in order to correctly account for the energetics of a variety of biological processes involving nucleic acids since they are highly charged polyelectrolytes. Due to the high charge density of nucleic acids, ions in the solution will condense around them. This “counterion cloud” that surrounds nucleic acids, which is not captured with structural techniques such as X-ray crystallography and NMR, is an integral part of their structure and essential to maintain their stability. Given this unique highly charged polyelectrolyte nature of nucleic acids it is not surprising that small changes in salt concentration can greatly affect its association with charged ligands (e.g., cationic drugs). Here, we examine the salt dependence of the association of more than 40 cationic minor groove antibiotics to AT-rich DNAs, for many of which thermodynamic binding data is available, using a novel implicit solvent-based Poisson-Boltzmann algorithm. We find that the plots of the electrostatic binding free energy versus the logarithm of salt concentration are linear, based on both the linear and nonlinear Poisson-Boltzmann equation. The slopes of these curves can be directly compared with experimental thermodynamic data of the observed binding constant at various salt concentrations. The good agreement between experimental and Poisson-Boltzmann predictions is only possible if the full nonlinear Poisson-Boltzmann equation is employed, since the linear PBE overestimates the slope of the plots of the electrostatic binding free energies vs. the logarithm of salt concentration. Thus, our results suggest that one should not employ the linear Poisson-Boltzmann or Generalized Born approach in order to assess the salt dependent behavior of nucleic acids. Further experimental and implicit solvent-based computational studies should be performed in order to verify if indeed the linear PBE always overestimates the salt dependence of the binding energetics of charged ligand-nucleic acid complexes.