Intercalative binding of ligands to DNA can be demonstrated by helix unwinding, monitored by gel electrophoresis of supercoiled DNA, as electrophoretic mobility is sensitive to the topological DNA state. However, we show that an apparent lack of unwinding in an electrophoretic assay could be due to dissociation of the (intercalated) ligand during the analysis, rather than evidence for a nonintercalative mode of binding to DNA. Repetitive scanning during the electrophoresis ensures that release of the ligand during electrophoresis does not affect the measured degree of unwinding, based on the electrophoretic velocity being determined as a function of time. We use this assay to establish intercalation as a mode of binding to DNA for the cyanine dyes YO, YO-PRO as well as two enantiomeric forms of the ruthenium complexes [(phen)2 Ru(tatpp)Ru(phen)(2)](4+), and to support groove-binding for the new unsymmetrical cyanine dyes BOXTO and BOXTO-PRO. Groove-binding could be concluded from a lack of unwinding, because we could rule out that it is caused by release of the dye during the electrophoresis. The gel electrophoresis has the advantage over hydrodynamic techniques that much smaller sample amounts are required, and our time-resolved approach can be employed in all mobility-shift assays when applied to dissociating complexes.