Single-cell proteo-genomic reveals a comprehensive map of centrosome-associated spliceosome components

Abstract

Ribonucleoprotein (RNP) condensates are crucial for controlling RNA metabolism and splicing events in animal cells. We used spatial proteomics and transcriptomic to elucidate RNP interaction networks at the centrosome, the main microtubule-organizing center in animal cells. We found a number of cell-type specific centrosome-associated spliceosome interactions localized in subcellular structures involved in nuclear division and ciliogenesis. A component of the nuclear spliceosome BUD31 was validated as an interactor of the centriolar satellite protein OFD1. Analysis of normal and disease cohorts identified the cholangiocarcinoma as target of centrosome-associated spliceosome alterations. Multiplexed single-cell fluorescent microscopy for the centriole linker CEP250 and spliceosome components including BCAS2, BUD31, SRSF2 and DHX35 recapitulated bioinformatic predictions on the centrosome-associated spliceosome components tissue-type specific composition. Collectively, centrosomes and cilia act as anchor for cell-type specific spliceosome components, and provide a helpful reference for explore cytoplasmic condensates functions in defining cell identity and in the origin of rare diseases.

Keywords: Biochemistry; Cell biology; Genomics; Proteomics.

Graphical abstract

Figure 1

The centrosome linker proteome network reflects the interactome size and complexity of eukaryotic organisms (A) Scheme of the protein interaction network involved in centrosome cohesion. (B) The protein interaction network of the centrosome linker proteins using the “BioGRID database”. (C) Left, the class and number of ribonucleoproteins interacting with centrosome linker proteins are shown graphically. Right, top GO terms of the CEP250 interactome (see Table S2). (D) Top, number of intrinsically disordered domains among centrosome linker proteins. Bottom, the higher number of intrinsically disordered domains in CEP250 protein correlates with the increased size of the interactome. Centrosome linker proteins are indicated by color circles. (E) Intrinsically disordered domains are correlated with the interactome of additional centrosome proteins including components of the pericentriolar material (PCNT), centriole growth (CEP350), and centrosome appendages (NIN, CEP164). (F) The levels of expression of centrosome linker, centriolar satellite, and centrosome appendage genes are absent or less represented in lower eukaryotes and non-mammalian organisms. PCM markers are shown for comparison.

Figure 2

Ribonucleoprotein particles interact with centrosomal proteins (A) Schematic representation of the centrosome with position and function of the nine bait proteins informed by Fogeron et al. (2013). (B) Left, the protein interaction maps for the centrosome baits and the translation and splicing components were analyzed according to SAINT (significance analysis of interactome) and are shown by dot plots. Dot shading (blue-black gradient) is the spectral counts for each prey protein. Dot size indicates relative abundance of prey protein in each analysis. Confidence levels and False discovery rate (FDR) are indicated by dot border. Right, the number of interactors for translation/splicing components related to centrosome substructures are shown graphically. (C) Mutations in both CEP250 and interacting splicing components are associated with inherited retinal blindness and microcephaly (green circles). (D) Spliceosome and translation complexes identified in different centrosome regions in neurons and neural stem cells informed by O'Neill et al. (2022).

Figure 3

Spliceosome components are localized at centrosomes and centriolar satellites (A) Schematic representation of the centrosome bait proteins according to Gupta et al. (2015). Centrosome proteins including components of the centriole, centriolar satellites and the cilium are indicated by different colors. (B) Left, protein interaction data are shown by SAINT. Right, the number of splicing interactors for each centrosome bait is indicated graphically. (C) Left, subcellular distribution of spliceosome components identified in the Human Protein Atlas database. Right, example of U251-MG cells stained with the splicing factor WDR83 localized at interphase centrosomes (white arrows) from the human proteome atlas. WDR83 (green), Alpha Tubulin (red) and nuclei (blue). (D) Subcellular distribution of ribonucleoproteins, eIFs, ribosomes (40S and 60S) and splicing components identified in different intracellular structures from the Human Protein Atlas database.

Figure 4

Splicing and centrosomal genes are regulated in a celltype-specific manner in humans and mice (A) Left, gene regulatory network of Fbf1 and Odf1 in mouse embryonic stem cells (mESC) according to the SINCERITIES algorithm. mESC data from Buettner et al. (2015). Right, the overlapping and divergent interactions with splicing components for Fbf1 and Ofb1 are shown by a Venn diagram. (B). Genetic interactions of different strengths between query (Fbf1 vs. Ofd1) and target splicing genes are shown by SAINT. (C) Left, the murine genetic network is recapitulated in the human centrosome protein interaction network. The major hub of this network includes BUD31 (5 interactions). Right, the interaction between OFD1 and BUD31 was confirmed through co-immunoprecipitation of OFD1 and verified by WB in HEK293 cells. Immunoprecipitated BUD31 was detected by WB. IgG were used as loading control. (D) A17 murine breast cancer stem cells stained with CEP250 (red), BCAS2 or SRSF2 (green). Nuclei are stained with Dapi in blue. Bar 50 μm. On the right of each panel: magnifications show the colocalization of CEP250/BCAS2 and CEP250/SRSF2 at centrosomes. Bar 10 μm.

Figure 5

Centrosome-associated splicing factors colocalize with CEP250 in fallopian tubes and cholangiocarcinoma (A) Left panel, IHC of CEP250, DHX35, BCAS2 and BUD31 in human fallopian tube (FT) tissues. Cilia are indicated by black arrows. Bar 50 μm. Right panel, top, subcellular distribution (Nuclear (N), cytoplasm (C) and membrane (M)) of each marker in ciliated and non-ciliated cells in 10 FT samples is shown graphically. Bottom, FT tissues stained with CEP250 (red), BCAS2 or BUD31 (green). Nuclei in blue. On the right of each panel: magnifications show the colocalization of CEP250/BCAS2 and CEP250/BUD31 at cilia. Bar 10 μm. (B) Left panel, IHC in human normal liver samples with markers used in A. Bar 50 μm. Right panel, top, staining intensity for all markers is reported as a boxplot. In the middle, the subcellular distribution (Nuclear (N), cytoplasm (C) and membrane (M)) for each marker in 7 samples is shown graphically. Bottom panel, FFPE tissue sample stained with CEP250 (red) and BCAS2 (green). Nuclei in blue. On the right: magnifications show that CEP250 and BCAS2 do not colocalize. Bar 10 μm. (C) Left panel, representative IHC images of CEP250 and BCAS2 staining in cholangiocarcinoma tissues. Right panel, cholangiocarcinoma sections stained with BCAS2 or BUD31 (green), CEP250 (red), and DNA (blue). On the right magnifications show that the colocalization of CEP250/BCAS2 and CEP250/BUD31 foci marks centrosomes in several cells. Bar 10 μm. (D) Quantification of the colocalization at centrosomes/cilia in the indicated tissues estimated according to Pearson’s correlation. From −1.0 to 1.0; 0 indicates no significant correlation and −1.0 indicates negative correlation. At least 100 cells in ten different high-magnifications fields (60x) tissue sections were examined. Statistical significance was evaluated using unpaired t tests. p values are reported in each graph.

Figure 6

Spliceosome components interact or reside at the centrosome or cilium (A) Schematic drawing showing the splicing proteins playing a role in the nuclear spliceosome that interact or localize at the centrosome. The function of splicing proteins and the temporal order of sub-complexes recruitment is given. (B) Schematic representation of the centrosome and the associated splicing components validated in this study.