Experiments establishing the intermediacy of the cumulene 3 and the biradical 4 in the cleavage of double-stranded DNA by neocarzinostatin chromophore (1) and methyl thioglycolate (2) are described. It is shown that, in the presence of millimolar concentrations of 2, greater than or equal to 95% of DNA cleavage arises via the cumulene 3; pathways of DNA cleavage not involving 3 are, at best, minor. The following detailed mechanism emerges for the DNA cleavage reaction. The rate-determining step in the damage of DNA by neocarzinostatin chromophore and the thiols glutathione (GSH), cysteine (CySH), or methyl thioglycolate at physiologically relevant concentrations and pH values is shown to be thiol addition to the chromophore. Evidence is presented to support the notion that the addition of thiols to 1, whether 1 is free in solution or bound to DNA, is an inherently efficient process. The addition of GSH or CySH is shown to proceed via a ternary complex of DNA, thiol, and chromophore. In neither case does DNA accelerate (catalyze) the thiol addition reaction; rather, it is found to induce a modest decrease in the rate of thiol addition versus control solutions lacking DNA. The greater concentration of DNA-bound chromophore versus chromophore free in solution offsets the attenuated rate of thiol addition to the former. The site of activation appears not to be critical, however, because it is likely that the cumulene intermediate produced is sufficiently long-lived to equilibrate among DNA binding sites. In support of this idea, it is shown that the sequence specificity of DNA cleavage by externally generated cumulene is identical to that of the presumptive cumulene formed in situ from 1 and 2 in the presence of DNA. It is proposed that the species that determines the sequence specificity of DNA cleavage is the cumulene intermediate. The experimental evidence suggests that the cumulene intermediate undergoes cycloaromatization while bound to DNA and that the biradical formed in this cycloaromatization reaction is a highly reactive and poorly selective intermediate. The yield-determining step in the production of thiol adducts from 1 is found to be the quenching of the biradical intermediate by hydrogen atom transfer. It is shown that double-stranded calf thymus DNA and the water-soluble 1,4-cyclohexadiene derivative 14 are approximately equally effective in trapping of the biradical intermediate at concentrations of 5 mM (base pairs) and 1 M, respectively, supporting the idea that the biradical is generated as a DNB-bound species. Although the data do not rule out the possibility that DNA may catalyze the cycloaromatization reaction, this proposal is considered to be unlikely. It is shown that thiol activation of 1 in the presence of single-stranded calf thymus DNA or a heterogeneous mixture of cellular RNA, but not bovine serum albumin, likely occurs as a ternary complex with the biopolymer. Furthermore, single-stranded DNA and heterogeneous cellular RNA are shown to serve as effective hydrogen atom donors for quenching of the biradical product of thiol activation, suggesting that biopolymers other than double-stranded DNA are potential targets for neocarzinostatin-induced damage.