The nuclear localization of the progesterone receptor is mediated by two signal sequences: one is constitutive and lies in the hinge region (between the DNA and steroid binding domains), the other is hormone dependent and is localized in the second zinc finger of the DNA binding domain. The use of various inhibitors of energy synthesis in cells expressing permanently or transiently the wild‐type receptor or a receptor mutated within the nuclear localization signals, demonstrated that the nuclear residency of the receptor reflects a dynamic situation: the receptor diffusing into the cytoplasm and being constantly and actively transported back into the nucleus. The existence of this nucleo‐cytoplasmic shuttle mechanism was confirmed by receptor transfer from one nucleus to the other in heterokaryons. Preliminary evidence was obtained, using oestrogen receptor, that this phenomenon may be of general significance for steroid receptors.

Deletion mutants of the rabbit progesterone receptor were used to identify two major mechanisms of its nuclear localization. A putative signal sequence, homologous to that of the SV40 large T antigen, was localized around amino acids 638-642 and shown to be constitutively active. When amino acids 638-642 were deleted, the receptor became cytoplasmic but could be shifted into the nucleus by the addition of hormone (or anti-hormone); it was almost fully active. The second mechanism consisted of the activation of the DNA binding domain. By deleting epitopes recognized by monoclonal antibodies, it was possible to follow different receptor mutants inside the same cells. In the absence of ligand, the receptor was transferred into the nucleus as a monomer. After administration of hormone (or anti-hormone) a "cytoplasmic" monomer was transferred into the nucleus through interaction with a "nuclear" monomer. These interactions occurred through the steroid binding domains of both monomers.

No abstract

Several nuclear proteins, including steroid hormone receptors, have been shown to shuttle continuously between the nucleus and the cytoplasm. The mechanism of entry of proteins into the nucleus is well documented, whereas the mechanism of their outward movement into the cytoplasm is not understood. We have grafted the nuclear llizton signals ofthe progesterone receptor or the simian virus 40 large tumor antigen onto fiJgalactosidase. These additions were shown to impart to the protein the ability to shuttle between the nucleus and the cytoplasm. Microinjected proteins devoid of a nuclear localization signal were unable to exit from the nucleus. The same nuclear localization signals are thus involved in both the inward and the outward movement of proteins through the nuclear membrane. We also show that although the nuclear import requires energy, the nuclear export does not. These results suggest that the nucleocytoplasmic shuttling may be a general phenomenon for nuclear proteins that could possibly undergo modifications in the cytoplasm and exert some biological activities there. These conclusions also imply that at least part of the celulr machinery involved in the nuclear import of proteins may finction bidirectionally.