A large amount of evidence suggests that arsenicals act via oxidative stress in causing cancer in humans and experimental animals. It is possible that arsenicals could bind in situ close to nuclear DNA followed by Haber-Weiss type oxidative DNA damage. Therefore, we tested this hypothesis by using radioactive (73)As labeled arsenite and vacuum filtration methodology to determine the binding affinity and capacity of (73)As arsenite to calf thymus DNA and Type 2A unfractionated histones, histone H3, H4 and horse spleen ferritin. Arsenicals are known to release redox active Fe from ferritin. At concentrations up to about 1 mM, neither DNA nor any of the three proteins studied, Type II-A histones, histone H3, H4 or ferritin, bound radioactive arsenite in a specific manner. Therefore, it appears highly unlikely that initial in situ binding of trivalent arsenicals, followed by in situ oxidative DNA damage, can account for arsenic's carcinogenicity. This experimental evidence (lack of arsenite binding to DNA, histone Type II-A and histone H3, H4) does not rule out other possible oxidative stress modes of action for arsenic such as (a) diffusion of longer lived oxidative stress molecules, such as H(2)O(2) into the nucleus and ensuing oxidative damage, (b) redox chemistry by unbound arsenicals in the nucleus, or (c) arsenical-induced perturbations in Fe, Cu or other metals which are already known to oxidize DNA in vitro and in vivo.