Selective FG phase in nuclear pores: lessons from engineering

Venue: Small Lecture Theatre, MSI

Host: Prof Dario Alessi

Abstract: The permeability barrier of nuclear pore complexes (NPCs) controls nucleocytoplasmic transport. It restricts inert macromolecules >30 kDa, but allows facilitated passage of nuclear transport receptor proteins (NTRs, e.g., importins and exportins) that shuttle cargoes in or out of the nucleus. The barrier can be described as a condensed protein-rich phase assembled from “cohesive” FG (Phe-Gly) repeat domains, which are long (~500 residues) intrinsically disordered regions at the central channel of NPCs. These FG domains include the most important Nup98 FG domain and several distinct subtypes, each comprising variable repeats and low complexity sequences.

The Nup98 FG domain comprises hydrophobic FG motifs, typically GLFG motifs, connected by more hydrophilic, uncharged spacers. These FG motifs bind NTRs passing NPCs. It was shown that the cohesive Nup98 FG domain phase-separates spontaneously from dilute aqueous solution to form gel-like phase (“FG phase”) with NPC-like selectivity: Such a condensed phase favours the partitioning of NTRs/NTR-cargo complexes, but excludes inert large proteins. However, the molecular details and determinants of this barrier property remained unclear. In particular, nanoscopic insight into the cohesive interactions has long been hampered by the sequence heterogeneity of the wild-type FG domain.

By systematically engineering the Nup98 FG domain, we identified several features in the sequence that are critical for the barrier property, for example, the frequency of FG motifs and the overall hydrophobicity of the domain. All of these correlate with the conserved features, suggesting that the FG domain is evolutionarily optimized for its barrier function. At one extreme of engineering, we obtained a variant composed of perfect repeats of 12 amino acid peptides, which still assembles into a barrier with exquisite transport selectivity. This barrier recapitulates importin- and exportin-mediated cargo transport and thus represents an ultimately simplified experimental model system. Such a system allowed us to overcome intrinsic challenges posed by the sequence heterogeneity of wild-type FG domains and to reveal the surprisingly fast local dynamics of cohesive interactions by high-resolution NMR spectroscopy on the FG phase.

The FG phase was initially thought to be an exotic form of biological matter. However, it is now clear that it is only a first example of a wide range of biomolecular condensates composed of sticky intrinsically disordered protein domains. My future work will involve investigating the regulatory mechanisms of these sticky domains, including how these regions are suppressed when they need to pass through the FG phase during nucleocytoplasmic transport.