It has recently been shown that large karyophilic proteins are transported across the nuclear envelope in amphibian oocytes. In consideration of this, the present experiments were performed to (a) identify the specific sites within the envelope through which transport occurs and (b) determine if molecular size is a limiting factor in the transport process. The following experimental procedure was employed: Colloidal gold particles, varying in size from ~20 to 170 A in diameter were coated with nucleoplasmin, a 165,000-mol-wt karyophilic protein, which is known to be transported through the envelope. The coated gold particles were microinjected into the cytoplasm of Xenopus oocytes, and the cells were fixed 15 min and 1 h later. The intracellular localization of the gold was then determined with the electron microscope. It was found that nucleoplasmin-coated particles readily enter the nucleus. On the basis of the distribution of the particles associated with the envelope, we concluded that transport occurs through the nuclear pores. Furthermore, the size distributions of the gold particles present in the nucleus and cytoplasm were not significantly different, indicating that the envelope does not discriminate among particles with diameters ranging from 50 to 200 ,g, (the dimensions including the nucleoplasmin coat). Colloidal gold coated with trypsin-digested nucleoplasmin (which lacks the polypeptide domain required for transport) or exogenous polyvinylpyrrolidone were largely excluded from the nucleus and showed no evidence of transport.Exogenous macromolecules appear to enter the nucleus by diffusing through central channels located within the nuclear pores (6). In amphibian oocytes it has been estimated that the channels available for passive diffusion are ~90 A in diameter (20). The rate of diffusion through these regions is inversely related to the size of the permeating substance. For example, in oocytes, exogenous molecules with hydrodynamic radii of ~ 15 A rapidly diffuse into the nucleus, whereas substances with radii of -45 ~, diffuse at extremely slow rates, if at all (20).Extrapolating to endogenous molecules, it is reasonable to assume that small proteins enter the nucleus by passive diffusion through the pores (although some form of facilitated uptake cannot be excluded). On the other hand, large nuclear proteins, which have a limited capacity for diffusion through 90 ~, pores, are presumably transported across the envelope. Evidence for transport has been obtained in two recent studies. Feldherr et al. (9) demonstrated that RN1, a 150,000-mol-wt nuclear protein found in Rana pipiens oocytes, enters the nucleus approximately 20 times faster than can be accounted for by diffusion through the pores. Dingwall et al. (3) showed that nucleoplasmin, a 165,000-mol-wt pentamer, is transported across the envelope. The transport of nucleoplasmin is prevented if a 12,000-mol-wt tail region, present on each of the monomeric subunits, is removed by protease digestion.This report is concerned mainly w...