Bound for the nucleus: defining the molecular principles of cargo selection by importin 9

Abstract

Active nuclear transport of proteins enables essential nuclear processes, such as genome packaging, transcription, splicing, ribosome biogenesis, and DNA repair. To facilitate proper sorting between cytoplasmic and nuclear compartments, the importin class of chaperone proteins transports hundreds of proteins through hydrophobic nuclear pores. While some importins recognize cargos via classical nuclear localization signals (cNLSs), there is not a comprehensive understanding of binding specificity determinants for those that appear to recognize other cargo features. Here we subject one such importin, IPO9, which is known to import H2A-H2B, TFIIB, actin, and the proteasome into the nucleus without cNLS binding, to a detailed analysis of its full set of cargos and the structural elements of IPO9 that confer their recognition. Through cytoplasmic immunoprecipitation followed by mass spectrometry (IP-MS), we stringently and reproducibly identify 79 bona fide IPO9-bound cargos, including 20 previously validated cargos. With this comprehensive cargo list, we find that IPO9 does not appear to use cNLSs, nor any other linear peptide motif, to identify and bind cargos. Unbiased oxidative footprinting of extracted IPO9*cargo complexes reveals that both the inner cavity of IPO9 and the unstructured loops protruding from the main body of the importin are protected by bound cargo, indicative of cargo competition for limiting IPO9 capacity. Guided by these data and evolutionary and phosphoproteomics insights, we employ a systematic IP-MS approach with loop perturbations to define how each element contributes to selective binding of subsets of the whole cargo cohort. These data suggest the H8 and H18-19 loops both specifically mediate IPO9 cargo-recognition by favorable enthalpic contacts. Additionally, these loops, as well as H7, appear to preclude binding of a secondary set of potential cargo to IPO9 which may normally be repelled by them or outcompeted by cognate cargo binding for which the loops provide attractive contacts. We define the continuum of cargo-release factor RanGTP sensitivity for our full cargo set, noting orders of magnitude range of sensitivity that argues that additional release factors may be necessary for efficient unloading in the nucleus. These experiments provide structure-function insight into importin binding specificity dictated by structural elements that recruit and/or restrict protein-protein interactions. Our approach of targeted mutations in cellular contexts coupled to quantitative proteomics affords a thorough biochemical dissection of the discrimination principles undergirding this unique molecular recognition problem of numerous-yet-specific binding events between importins and their many distinct cargos.