Ku antigen is a complex of Ku70 and Ku80 subunits and plays an important role in not only DNA double-strand breaks (DSB) repair and V(D)J recombination, but also in growth regulation. Ku is generally believed to always form and function as heterodimers on the basis of in vitro observations. Here we demonstrate that the localization of Ku80 does not completely coincide with that of Ku70. Ku70 and Ku80 were colocalized in the nucleus in the interphase but not in the late telophase/early G1 phase of the cell cycle. Since the in vivo function of Ku might be partially regulated by the control of its transport, we attempted to investigate the molecular mechanisms underlying the nuclear translocation of Ku. The nuclear translocation of Ku80 started during the late telophase/ early G1 phase after the nuclear envelope was formed and this was preceded by the nuclear translocation of Ku70. Furthermore, we found that the Ku80 protein was transported to the nucleus without heterodimerization with Ku70. To understand in detail the mechanism of transport of Ku80, we attempted to identify the nuclear localization signal (NLS) of Ku80 and de®ned to a region spanning nine amino acid residues (positions 561 ± 569). The Ku80 NLS was demonstrated to be mediated to the nuclear rim by two components of PTAC58 and PTAC97. All these ®ndings support the idea that Ku80 can translocate to the nucleus using its own NLS independent of the translocation of Ku70.