Protein localization to the nucleolus: a search for targeting domains in nucleolin.

Nucleolin, a major nucleolar phosphoprotein, is presumed to function in rDNA transcription, rRNA packaging and ribosome assembly. Its primary sequence was highly conserved during evolution and suggests a multi-domain structure. To identify structural elements required for nuclear uptake and nucleolar accumulation of nucleolin, we used site-directed mutagenesis to introduce point- and deletion-mutations into a chicken nucleolin cDNA. Following transient expression in mammalian cells, the intracellular distribution of the corresponding wild-type and mutant proteins was determined by indirect immunofluorescence microscopy. We found that nucleolin contains a functional nuclear localization signal (KRKKEMANKSAPEAKKKK) that conforms exactly to the consensus proposed recently for a bipartite signal (Robbins, J., Dilworth, S.M., Laskey, R.A. and Dingwall, C. (1991) Cell 64, 615-623). Concerning nucleolar localization, we found that the N-terminal 250 amino acids of nucleolin are dispensible, but deletion of either the centrally located RNA-binding motifs (the RNP domain) or the glycine/arginine-rich C terminus (the GR domain) resulted in an exclusively nucleoplasmic distribution. Although both of these latter domains were required for correct subcellular localization of nucleolin, they were not sufficient to target non-nucleolar proteins to the nucleolus. From these results we conclude that nucleolin does not contain a single, linear nucleolar targeting signal. Instead, we propose that the protein uses a bipartite NLS to enter the nucleus and then accumulates within the nucleolus by virtue of binding to other nucleolar components (probably rRNA) via its RNP and GR domains.