Nucleolar URB1 ensures 3' ETS rRNA removal to prevent exosome surveillance.

The nucleolus is the most prominent membraneless condensate in the nucleus. It comprises hundreds of proteins with distinct roles in the rapid transcription of ribosomal RNA (rRNA) and efficient processing within units comprising a fibrillar centre and a dense fibrillar component and ribosome assembly in a granular component1. The precise localization of most nucleolar proteins and whether their specific localization contributes to the radial flux of pre-rRNA processing have remained unknown owing to insufficient resolution in imaging studies2-5. Therefore, how these nucleolar proteins are functionally coordinated with stepwise pre-rRNA processing requires further investigation. Here we screened 200 candidate nucleolar proteins using high-resolution live-cell microscopy and identified 12 proteins that are enriched towards the periphery of the dense fibrillar component (PDFC). Among these proteins, unhealthy ribosome biogenesis 1 (URB1) is a static, nucleolar protein that ensures 3' end pre-rRNA anchoring and folding for U8 small nucleolar RNA recognition and the subsequent removal of the 3' external transcribed spacer (ETS) at the dense fibrillar component-PDFC boundary. URB1 depletion leads to a disrupted PDFC, uncontrolled pre-rRNA movement, altered pre-rRNA conformation and retention of the 3' ETS. These aberrant 3' ETS-attached pre-rRNA intermediates activate exosome-dependent nucleolar surveillance, resulting in decreased 28S rRNA production, head malformations in zebrafish and delayed embryonic development in mice. This study provides insight into functional sub-nucleolar organization and identifies a physiologically essential step in rRNA maturation that requires the static protein URB1 in the phase-separated nucleolus.

lncRNA SLERT controls phase separation of FC/DFCs to facilitate Pol I transcription.

RNA polymerase I (Pol I) transcription takes place at the border of the fibrillar center (FC) and the dense fibrillar component (DFC) in the nucleolus. Here, we report that individual spherical FC/DFC units are coated by the DEAD-box RNA helicase DDX21 in human cells. The long noncoding RNA (lncRNA) SLERT binds to DDX21 RecA domains to promote DDX21 to adopt a closed conformation at a substoichiometric ratio through a molecular chaperone-like mechanism resulting in the formation of hypomultimerized and loose DDX21 clusters that coat DFCs, which is required for proper FC/DFC liquidity and Pol I processivity. Our results suggest that SLERT is an RNA regulator that controls the biophysical properties of FC/DFCs and thus ribosomal RNA production.

An optimized fixation method containing glyoxal and paraformaldehyde for imaging nuclear bodies.

The mammalian cell nucleus contains different types of membrane-less nuclear bodies (NBs) consisting of proteins and RNAs. Microscopic imaging has been widely applied to study the organization and structure of NBs. However, current fixation methods are not optimized for such imaging: When a fixation method is chosen to maximize the quality of the RNA fluorescence in situ hybridization (FISH), it often limits the labeling efficiency of proteins or affects the ultrastructure of NBs. Here, we report that addition of glyoxal (GO) into the classical paraformaldehyde (PFA) fixation step not only improves FISH signals for RNAs in NBs via augmented permeability of the fixed nucleus and enhanced accessibility of probes, but also largely preserves protein fluorescent signals during fixation and immunostaining. We also show that GO/PFA fixation enables the covisualization of different types of nuclear bodies with minimal impact on their ultrastructures under super-resolution microscopy.

Dynamic Imaging of RNA in Living Cells by CRISPR-Cas13 Systems.

Visualizing the location and dynamics of RNAs in live cells is key to understanding their function. Here, we identify two endonuclease-deficient, single-component programmable RNA-guided and RNA-targeting Cas13 RNases (dCas13s) that allow robust real-time imaging and tracking of RNAs in live cells, even when using single 20- to 27-nt-long guide RNAs. Compared to the aptamer-based MS2-MCP strategy, an optimized dCas13 system is user friendly, does not require genetic manipulation, and achieves comparable RNA-labeling efficiency. We demonstrate that the dCas13 system is capable of labeling NEAT1, SatIII, MUC4, and GCN4 RNAs and allows the study of paraspeckle-associated NEAT1 dynamics. Applying orthogonal dCas13 proteins or combining dCas13 and MS2-MCP allows dual-color imaging of RNAs in single cells. Further combination of dCas13 and dCas9 systems allows simultaneous visualization of genomic DNA and RNA transcripts in living cells.

Nascent Pre-rRNA Sorting via Phase Separation Drives the Assembly of Dense Fibrillar Components in the Human Nucleolus.

Fibrillar centers (FCs) and dense fibrillar components (DFCs) are essential morphologically distinct sub-regions of mammalian cell nucleoli for rDNA transcription and pre-rRNA processing. Here, we report that a human nucleolus consists of several dozen FC/DFC units, each containing 2-3 transcriptionally active rDNAs at the FC/DFC border. Pre-rRNA processing factors, such as fibrillarin (FBL), form 18-24 clusters that further assemble into the DFC surrounding the FC. Mechanistically, the 5' end of nascent 47S pre-rRNA binds co-transcriptionally to the RNA-binding domain of FBL. FBL diffuses to the DFC, where local self-association via its glycine- and arginine-rich (GAR) domain forms phase-separated clusters to immobilize FBL-interacting pre-rRNA, thus promoting directional traffic of nascent pre-rRNA while facilitating pre-rRNA processing and DFC formation. These results unveil FC/DFC ultrastructures in nucleoli and suggest a conceptual framework for considering nascent RNA sorting using multivalent interactions of their binding proteins.