Cell. Vol. 56, 379-390, February 10, 1989, Copyright 0 1969 by Cell Press Major Nucleolar Proteins Shuttle between Nucleus and Cytoplasm R. A. Borer,’ C. F. Lehner,t H. M. Eppenberger, and E. A. Nigg* * lnstitut fur Zellbiologie ETH Hdnggerberg CH-8093 Zurich Switzerland tDept. of Biochemistry and Biophysics University of California at San Francisco San Francisco, California 94143 *Swiss Institute for Experimental Cancer Research Chemin des Boveresses 155 CH-1066 Epalinges Switzerland Summary Nucleolin is a 92 kd nucleolar protein implicated in regulating polymerase I transcription and binding of preribosomal RNA. Another abundant nucleolar pro- tein of 38 kd (B23/No38) is thought to be involved in intranuclear packaging of preribosomal particles. Al- though both proteins have previously been detected only in nuclei, we conclude that they shuttle con- stantly between nucleus and cytoplasm. This conclu- sion is based on monitoring the equilibration of these proteins between nuclei present in interspecies het- erokaryons, and on observing the antigen-mediated nuclear accumulation of cytoplasmically injected anti- bodies. Our unexpected results suggest a role for these major nucleolar proteins in the nucleocytoplas- mic transport of ribosomal components. Moreover, they suggest that transient exposure of shuttling pro- teins to the cytoplasm may provide a mechanism for cytoplasmic regulation of nuclear activities. introduction The existence of nuclear proteins that shuttle back and forth between nucleus and cytoplasm was first inferred from nuclear transplantation studies carried out on amoe- bae (Goldstein, 1958; for review, see Goldstein and Ko, 1981). Unfortunately, the shuttling amoebae proteins have not been functionally characterized, and, for technical reasons, it was difficult to explore the generality of the phenomenon in other organisms. Nucleocytoplasmic shuttling of proteins in higher eukaryotic cells has been suggested on the basis of cell fusion experiments (e.g., Appels and Ringertz, 1975; Woodcock et al., 1984) and microinjection studies (Bonner, 1975; Madsen et al., 1986) but alternative explanations have not been rigorously excluded. Moreover, with the exception of the major nonhistone protein HMGl (Rechsteiner and Kuehl, 1979) none of the purported shuttling proteins has been identified. Here we have used two independent approaches for studying the shuttling of individual nuclear proteins be- tween nucleus and cytoplasm (outlined in Figure 1). The first type of assay consisted of monitoring thle migration of (chicken) nuclear proteins in interspecies (chick-mouse) heterokaryons: in these experiments the mouse nucleus was expected to act as a trap for any chicken nuclear pro- tein that appeared, at least temporarily, in the cytoplasm of the heterokaryon. That nuclear proteins originating from one vertebrate species would enter nuclei from other species was suggested by earlier studies (e.g., Appels and Ringertz, 1975; Jost et al., 1979; Scheer et al., 1983; Nyman et al., 1984; Woodcock et al., 1984). The second type of experiment consisted of observing the distribution of antibodies following their injection into the cytoplasm of cultured (chicken) cells. This approach was lbased on the observation that immunoglobulins (IgG) canrnot passively diffuse through nuclear pores (Stacey and Allfrey, 1984; Einck and Bustin, 1984). Therefore, cytoplasmically in- jected antibodies were expected to remain in the cyto- plasm, unless they crossed the nuclear envelope in the form of complexes with their corresponding nuclear anti- gens (Sugawa et al., 1985; Tsuneoka et al., 1986; Madsen et al., 1986). In such cases, nuclear uptake of antigen- antibody complexes was presumed to be Imediated by karyophilic signal sequences present on the antigens (for recent reviews on nuclear transport, see Dingwall and Laskey, 1986; Peters, 1986; Newport and Forbes, 1987). One of the most fundamental cellular processes requir- ing extensive transport of proteins and ribonucleoprotein particles across the nuclear envelope is riboslome biogen- esis (for review, see Hadjiolov, 1985; Sommerville, 1986; Nigg, 1988). Synthesis of the 45s rRNA preclursor and its stepwise processing to 28S, 18S, and 5.8s mature ribosomal RNAs occur in the nucleolus, whereas synthe- sis of 55 RNA takes place on extranucleofar chromatin. Fiibosomal proteins are translated in the cytoplasm, and therefore need to be imported into the nucleus before they can be assembled with ribosomal RNAs tlo form pre- ribosomal particles. Nonribosomal nucleolar proteins re- quired for processing and packaging of the,se part/cles are generally believed to dissociate from ‘the maturing ribosomal subunits before these are exported to the cytoplasm. Our present results lead us to conclude that two non- ribosomal nucleolar proteins of 92 kd and 38 kd, respec- tively, migrate constantly back and forth between nucleus and cytoplasm; accordingly, we shall refer :to these pro- teins as shuttling proteins. Based on partia’l sequencing of cDNA clones, we have identified the 92 kd protein as the chicken homolog of the major mammalian nucleolar protein known as nucleolin (Lapeyre et al., ‘1987) or C23 (Busch, 1984). Likewise, the 38 kd protein ‘was found to’be homologous to the mammalian protein 823 (Busch, 1984; Chan et al., 1986a) and to the amphibian nuc!eolar protein No38 (Schmidt-Zachmann et al., 1987). Nucleolin/C23 is believed to be a multifunctional’protein (for ireview, see Jordan, 1987) and hasbeen implicated in the control of transcription of ribosomal RNA genes by RN,A polymer-