The biochemical characteristics of the formation and disappearance of intercalator-induced DNA double-strand breaks (DSB) were studied in nuclei from mouse leukemia L1210 cells by using filter elution methodology. The 3 intercalators used were 4''-(9-acridinylamino)-methanesulfon-m-anisidide (m-AMSA), 5-iminodaunorubicin (5-ID) and ellipticine. These compounds differ in that they produced predominantly DNA single-strand breaks (SSB) (m-AMSA) or predominantly DNA double-strand breaks (ellipticine) or a mixture of both SSB and DSB (5-ID) in whole cells. In isolated nuclei, each intercalator produced DSB at a frequency comparable to that which is produced in whole cells; these DNA breaks reversed within 30 min after drug removal. It thus appeared that neither ATP nor other nucleotides were necessary for intercalator-dependent DNA nicking-closing reactions. The formation of the intercalator-induced DSB was reduced at ice temperature. Break formation was also reduced in the absence of Mg, at a pH > 6.4 and at NaCl concentrations > 200 mM. In the presence of ATP and ATP analogs, the intercalator-induced cleavage was enhanced. The intercalator-induced DSB are probably enzymatically mediated. The enzymes involved in these reactions can probably catalyze DNA double-strand cleavage and rejoining in the absence of ATP, although the occupancy of an ATP binding site might convert the enzyme to a form more reactive to intercalators. Three inhibitors of DNA topoisomerase II (novobiocin, nalidixic acid and norfloxacin) reduced the formation of DNA strand breaks. These findings are consistent with the hypothesis that intercalator-induced DNA breakage results from the DNA cleavage action of a DNA topoisomerase II.