Fig 2: Plk1 inhibition in late G2 allows centriole duplication in post-mitotic cells.(a) Analysis of centrosome-associated Sas6 signal by immunofluorescence. In control G1 cells, mother and daughter centrioles are separated and Sas6 is absent. In the 4hBIRO sample, mother and daughter centrioles are adjacent, and associated with one Sas6. 1hBIRO sample contains two adjacent centrioles associated with two Sas6 signals, indicative that mother centriole is reduplicating. Schematics of putative centriole configurations are provided to facilitate understanding. Histogram represents quantification of the cells with 0, 2 or >2 Sas6 signals. Average value and the s.d. are presented. n=300, from three independent experiments. (b) In 1hBIRO cells, one of the Sas6 signals is associated with a mother centriole (brighter C1–GFP signal) and is co-localized with Plk4 signal. (c) In centrosomes with two Sas6 signals, the older daughter centriole is organizing gamma tubulin and Cep152. Its Sas6 signal is positioned outside the Cep152 toroid organized by the mother centriole. The centre of nascent Sas6 signal co-localizes with the mother's Cep152 toroid. Scale bars, 400 nm (a–c). (d) Correlative time-lapse and electron microscopy analysis of a post-mitotic cell from 1hBIRO samples. A 5-min-long time-lapse recording with 10 s resolution reveals weak C1–GFP signals (arrows) associated with the mother centrioles, in addition to brighter C1–GFP signals. Electron microscopy further confirmed that the weak C1–GFP signals correspond to the nascent daughter centrioles (D1–2 and D2–2) positioned close to the wall of the mother centrioles. Original daughter centrioles (D1 and D2) were still in an orthogonal orientation to the mother centriole. Scale bars, 200 nm.

Fig 3: Analysis of ultra-structural changes within the centrosome during the earliest stages of disengagement.(a,b) Centrioles of C1–GFP-expressing cells arrested in G2 by Cdk1 inhibitor RO-3306 (RO) were first analysed by 5 min time lapse to determine the pattern of centriole movement (expressed as the average distance between the mother and daughter C1–GFP signals). The same centrioles were then fixed and analysed by electron microscopy. Mother–daughter centriole pairs with average C1–GFP distance ≤700 nm were found in orthogonal orientation and at the wall-to-wall distance <80 nm. Daughter centrioles ≥80 nm from the wall of the mother centrioles were found disoriented. Scale bars, 500 nm. (c,d) Daughter centrioles in the centrosomes with C1–GFP distances ≤700 nm are associated with a substantial amount of PCM (as evidenced by a gamma tubulin signal) and with licensing factor Cep152. Accumulation of Cep152 precedes the degradation of Sas6, separation of two PCMs and loss of orthogonal orientation between the centrioles. Scale bar, 1 μm. (e) Quantification of daughter centriole-associated gamma tubulin and Cep152. Only centrioles with mother centriole perpendicular to the coverslip were measured. For each centrosome, two signal lengths were measured from the centre of a gamma tubulin or Cep152 toroid organized by the mother centriole, until the end of the signal associated with the daughter centriole (determined as a signal half maximum). Centrioles in cells arrested in G2 and treated with Plk1 inhibitor BI2536 (RO+BI) were measured as a control.

Fig 4: Effect of depletion of STIL or HsSAS6 on the behavior of centriolar Plk4. (A) Time course of centriolar Plk4-mClover fluorescence in cells transfected with siControl (top), siSTIL (mid) or siHsSAS6 (bottom). The results in the graphs are shown as mean±s.d. of 8, 10 and 8 cells, respectively. (B,C) Effect of depletion of STIL or HsSAS6 on the spatial patterning of centriolar Plk4. Representative high-resolution immunofluorescence images of Plk4 with and without the centriole marker CEP152 (B) and the ring-filling indices (C) for the three conditions. Scale bar: 0.5 μm. (D) Interdependency of centriolar accumulation of STIL and HsSAS6. HeLa cells were transfected with siControl, siSTIL or siHsSAS6 and then stained with antibodies to STIL or HsSAS6 and the centriole marker pericentrin (PCNT). Representative images are shown for each combination. Scale bar: 1 μm. A.U., arbitrary units.

Fig 5: Asymmetrical distribution of Cep152 in RPE-1 cells and its relationship with Cep192. Logarithmically growing RPE-1C1-GFP cells were immunolabeled for Cep44, Cep192, Cep57, Cep63, or Cep152. Cep152 was detected using an antibody against its N-terminus (Nt) or middle portion (Mid; Fig. S1 A). Some samples were additionally expanded twofold and labeled for acetylated tubulin. (A) The discontinuous appearance of Cep152 signals with gaps (arrows) extending over a variable number of centriole microtubule blades. Centrioles are shown in the top view. (B) The discontinuous appearance of Cep57 and Cep63 showing gaps (arrows). In contrast, Cep44 and Cep192 show continuous ring-like distribution. (C) Cep57, Cep63, and Cep152 signal in relation to Centrin1-GFP signal marking centrioles. Regardless of the gaps in Cep57, Cep63, and Cep152 signals, mother centrioles duplicate and regularly form only one daughter centriole. (D) Quantification of centrosomes with gaps in the signal of the indicated protein around the mother centriole. Centrioles imaged in vertical or near vertical orientation were quantified. (E and F) RPE-1Plk4 cells were treated with Dox for 18 h to induce overexpression of FLAG-Plk4 and formation of centriolar rosettes, and cells were immunolabeled for indicated proteins. (E) Centrosomes with centriolar rosettes to illustrate the absence of procentrioles at regions lacking Cep57, Cep63, and Cep152. Cep192 forms a complete ring around mother centrioles. Right: The absence of Plk4 signal on the regions of the centrosome lacking Cep152. (F) Quantification of centrosomes in RPE-1Plk4 cells exhibiting gaps in the signals of Cep192, Cep57, Cep63, and Cep152. Centrioles imaged in vertical or near vertical orientation were quantified. (G) Experimental strategy used in H–K (for more details, please see Materials and methods). (H) The effect of Cep192 or Cep152 depletion on the assembly of centriolar rosettes. Cells were depleted for Cep192 or Cep152, and centriolar rosettes were induced by transient treatment with CEN and Cep192 and Cep152 were immunolabeled. Centrin1-GFP marks centrioles. (I) Quantification of cells from H containing centriolar rosettes, duplicated centrioles, or at least one unduplicated centriole. 200 cells were counted per condition in each experiment. (J) The effect of Cep192 or Cep152 depletion on Cep152 and Cep192 centrosomal levels. Intensities of Cep192 and Cep152 were determined from widefield recordings and plotted for individual centrosomes. (K) The effect of Cep192 depletion on centrosomal Plk4 levels. Plk4 and Cep192 were labeled and their centrosomal intensity was quantified from widefiled images. Red lines: average intensity. (L) Localization of Cep57, Cep192, Cep63, Cep152, and SAS-6 on de novo–formed centrioles in S phase RPE-1Plk4 cells treated with Dox for 18 h. The wider end of the SAS-6 signal distinguishes the centriole’s proximal (p) end from the distal (d) end. Cep192 associates with the centrioles laterally, Cep63 and Cep152 associate with the centriole’s proximal ends, while Cep57 does not associate with de novo formed centrioles. Scale bars: 0.5 µm for STED, 1 µm for twofold expansion/STED, and 2 µm for widefield.