Centriolar satellites are acentriolar assemblies of centrosomal proteins

Abstract Centrioles are core structural elements of both centrosomes and cilia. Although cytoplasmic granules called centriolar satellites have been observed around these structures, lack of a comprehensive inventory of satellite proteins impedes our understanding of their ancestry. To address this, we performed mass spectrometry (MS)‐based proteome profiling of centriolar satellites obtained by affinity purification of their key constituent, PCM1, from sucrose gradient fractions. We defined an interactome consisting of 223 proteins, which showed striking enrichment in centrosome components. The proteome also contained new structural and regulatory factors with roles in ciliogenesis. Quantitative MS on whole‐cell and centriolar satellite proteomes of acentriolar cells was performed to reveal dependencies of satellite composition on intact centrosomes. Although most components remained associated with PCM1 in acentriolar cells, reduced cytoplasmic and satellite levels were observed for a subset of centrosomal proteins. These results demonstrate that centriolar satellites and centrosomes form independently but share a substantial fraction of their proteomes. Dynamic exchange of proteins between these organelles could facilitate their adaptation to changing cellular environments during development, stress response and tissue homeostasis.


*
. The effect of combined nocodazole and cytochalasin-B treatment on the distribution of endogenously labelled PCM1-GFP in DT40 cells.
A Diagram showing the GFP construct used to target the chicken PCM1 locus at the C-terminus on both alleles, by homologous recombination. Highlighted the Sal1 and BamH1 sites used for restriction digestion to clone the LA (Left Arm) and the RA (Right Arm) and to replace the resistance cassette. Clones were screened for antibiotic resistance genes blasticidin (Blasti), puromycin (Puro) or Histidinol (His). LoxP sites flanking the resistance cassette are represented by red triangles. The dashed lines indicate the sites of recombination and integration in the PCM1 locus. Confirmation of targeting was carried out by Western blotting, as shown in Fig 1B-D. B Representative immunofluorescence images of cell lines with genotypes as indicated, treated with both nocodazole (2 lg/ml) and cytochalasin-B (1 lg/ml). DMSO-treated cells were used as a control (DMSO, upper panels). Treatments were carried out for 2 h, and cells were co-stained with antibodies against GFP (green) and c-tubulin (red). DNA is in blue. Images correspond to maximum intensity projections of confocal micrographs. Asterisks mark cells with dispersed satellites. Note that drug treatment leads to an increase in large and a decrease in small satellite granules in all three genotypes, but the effects are more prominent in acentriolar than in WT cells. Scale bars: 5 lm.
▸ Figure EV2. Comparisons of CS-WT with published datasets.
A Venn diagram showing the number of proteins identified in three datasets: CS-WT, PCM1-BioID (Gupta et al, 2015) and PCM1-FLAG IP (Gupta et al, 2015). Note that this and all subsequent analyses were performed on human orthologues of the chicken proteins from CS-WT. B Venn diagrams showing the number of published CS components (illustrated in the table in C) detected in each of the indicated datasets. C Table depicts    A Representative immunofluorescence images of Jurkat cells co-stained with antibodies against selected new CS candidates (green) and PCM1 (red). SSX2IP, a known CS component, is shown as positive control. The framed panel at the bottom illustrates the relative distributions of the centrosomal marker c-tubulin (red) and the CS protein PCM1 (green) in Jurkat cells. DNA is in blue. Images correspond to maximum intensity projections of confocal micrographs. High magnification images are included to aid visualisation of framed areas. Scale bars: 5 lm. B Representative immunofluorescence images of Jurkat cells mock-treated with DMSO or incubated with nocodazole (2 lg/ml) to depolymerise microtubules. Cells were co-stained with antibodies against PCM1 (green), c-tubulin (red) and a-tubulin (blue). Nocodazole reduces PCM1 signal in the pericentrosomal region. High magnification images are included to aid visualisation of framed areas and correspond to framed areas. Scale bar: 5 lm. C Knock-down efficiency of siRNAs assessed by qPCR. The relative expression of each candidate upon siRNA treatment was assessed by qPCR and shown relative to cells treated with a control siRNA. Each datapoint represents a biological replicate. Note that T3JAMsi1 enhances rather than reduces mRNA expression. Bar graphs show mean AE SE. -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 -6 -2 2 6 A CRISPR/Cas9 strategy to generate PCM-deficient cells. The PAM sequences are highlighted in bold. Target sequences in exons 3 and 26 are shown. Below, predicted translational products from all the sequenced variants are listed. Briefly, PCM1 variants were identified by amplifying and sequencing the regions of PCM1 cDNA targeted by the gRNAs. Amino acids deleted are represented by the dash symbol, those divergent from the control sequence are highlighted in red, whereas stop of translation is represented by the asterisk. For each clone, 10-25 bacterial colonies were sequenced. The percentage of each transcript variant/translational product is indicated at the right of each panel. For PCM1-KO 3, the PCM1 cDNA sequence was not detected. B Representative immunofluorescence images of PCM1 control (CON 1) and KO (KO 3) cells co-stained with antibodies against PCM1 (green) and c-tubulin (red). DNA is in blue. Images correspond to maximum intensity projections of wide-field micrograph. Scale bar: 10 lm. In unfiltered data In filtered data In CS-WT Figure EV6. Centrosomal abundance of proteins positively correlates with their presence in the satellite proteome.
A Graph depicts relative protein abundance in centrosome preparations from KE37 cells (adapted from Fig 3, Bauer et al, 2016) and their status in our datasets. The categories Unfiltered and Filtered Data and CS-WT are defined in Fig 2A. Proteins absent from the chicken genome have been removed. See also Table EV4. B To assay correlation between centrosomal abundance of proteins (from Bauer et al, 2016) and their presence in our satellite proteome datasets, centrosomal proteins were categorised based on their detection in our satellite proteome datasets. For each category (presence vs. absence), the abundance ranks of protein copy numbers were then compared using a two-sided Wilcoxon rank-sum test and the corresponding P values are depicted in the table.