Competition between targeting signals in hybrid proteins provides information on their relative in vivo affinities for subcellular compartments

• Published in: European journal of cell biology 2008-02, Vol.87 (2), p.57-68
• Main Authors: Schmauch, Christian, Maniak, Markus
• Format: Article
• Language: English
• Subjects: Actin; Animals; Cytoplasm - metabolism; Cytoskeletal Proteins - metabolism; Cytoskeleton; Cytoskeleton - metabolism; Dictyostelium; Dictyostelium - metabolism; Endosome; Endosomes - metabolism; Flotillin; Lysosome; Microfilament Proteins - metabolism; Nuclear Localization Signals - metabolism; Nucleus; Peroxisome; Protein Binding; Protein Sorting Signals - physiology; Protein Transport; Protozoan Proteins - metabolism; Recombinant Fusion Proteins - metabolism; Reggie; Signal sequence; Signal Transduction
• ISSN: 0171-9335; 1618-1298
• DOI: 10.1016/j.ejcb.2007.10.003

After their translation and folding in the cytoplasm, proteins may be imported into an organelle, associate with a membrane, or rather become part of large, highly localised cytoplasmic structures such as the cytoskeleton. The localisation of a protein is governed by the strength of binding to its immediate target, such as an import receptor for an organelle or a major component of the cytoskeleton, e.g. actin. We have experimentally provided a set of actin-binding proteins with competing targeting information and expressed them at various concentrations to analyse the strength of the signal that governs their subcellular localisation. Our microscopic observations indicate that organellar sorting signals override the targeting preference of most cytoskeletal proteins. Among these signals, the nuclear localisation signal of SV40 is strongest, followed by the oligomerised PHB domain that targets vacuolin to the endosomal surface, and finally the tripeptide SKL mediating transport into the peroxisome. The actin-associated protein coronin, however, can only be misled by the nuclear localisation signal. Interestingly, the targeting behaviour of this model set of hybrid proteins in living Dictyostelium amoebae correlates surprisingly well with the affinities of their constituent signals derived from in vitro experiments conducted in various other organisms. Accordingly, this approach allows estimating the in vivo affinity of a protein to its target even if the latter is not known, as in the case of vacuolin.