Rab11 endosomes and Pericentrin coordinate centrosome movement during pre-abscission in vivo

Abstract

The last stage of cell division involves two daughter cells remaining interconnected by a cytokinetic bridge that is cleaved during abscission. Conserved between the zebrafish embryo and human cells, we found that the oldest centrosome moves in a Rab11-dependent manner towards the cytokinetic bridge sometimes followed by the youngest. Rab11-endosomes are organized in a Rab11-GTP dependent manner at the mother centriole during pre-abscission, with Rab11 endosomes at the oldest centrosome being more mobile compared with the youngest. The GTPase activity of Rab11 is necessary for the centrosome protein, Pericentrin, to be enriched at the centrosome. Reduction in Pericentrin expression or optogenetic disruption of Rab11-endosome function inhibited both centrosome movement towards the cytokinetic bridge and abscission, resulting in daughter cells prone to being binucleated and/or having supernumerary centrosomes. These studies suggest that Rab11-endosomes contribute to centrosome function during pre-abscission by regulating Pericentrin organization resulting in appropriate centrosome movement towards the cytokinetic bridge and subsequent abscission.

Figure 1.. Differences in mitotic centrosome movement towards the cytokinetic bridge during pre-abscission between zebrafish embryos and human cells.

(A) Zebrafish embryo (5 h post fertilization) with centrin-GFP (gray) and PLK1-mCh (cyan). Scale bar, 50 μm. (a’) Inset of dividing cell in (A). Time-lapse of centrin-GFP (inverted grays, top panel; grays, bottom panel) and PLK1-mCh (cyan). Video 1 Pink arrow, centrosome. Orange arrow, midbody. Dashed lines, cell boundaries. Scale bar, 10 μm. (B) Model depicting centrosome (green) movement towards the cytokinetic bridge in dividing cells within the Kupffer’s vesicle (KV) during its development. Cyan, Nucleus. Purple, Midbody. Orange, Lumen. Dark lines, KV membranes. (C) Time-lapse of a dividing cell within the KV. KV cell membranes marked with Sox17:GFP-CAAX (gray). Cyan star, rosette center. Cyan arrow, centrosome. Dashed lines, cell boundaries. Scale Bar, 10 μm. (c’) Dividing cell depicted with PLK1-mCh (fire LUT). Cyan arrow, centrosome. (D) A KV pre-abscising cell fixed and immunostained for ZO-1 (gray), γ-tubulin (cyan) and DNA (DAPI, blue). Yellow arrow, centrosome. Dashed lines, cell boundaries. Scale bar, 10 μm. (E) Number of centrosomes per pre-abscising cell with bridge directed centrosome movement calculated as both centrosomes (2 centrosomes), only one centrosome (1 centrosome) and neither centrosome (0 centrosomes) moved shown as a violin plot with median (orange) and quartiles (dark dotted lines). Two-tailed t test between Centrin-GFP and DsRed-PACT in Human (HeLa) cells (pink background). n > 10 cells across n > 3 experiments n.s. not significant. One-way ANOVA across zebrafish epiboly cells and KV cells (green background), n.s. not significant. n > 10 cells across n > 2 embryos. One-way ANOVA, across all columns, **P < 0.01. Two-tailed t test between Human (HeLa) cells and zebrafish (Epiboly, KV) cells, **P < 0.01. n-values, detailed statistical results in Table S1.

Figure S1.. Differences in mitotic centrosome movement towards the cytokinetic bridge during pre-abscission between zebrafish embryos and human cells.

(A, B) Overlay of centrosome movement within pre-abscising zebrafish embryo cell (A) and human (HeLa) cell (B) over time. The centrosome is labeled with centrin-GFP (zebrafish embryo cell, gray) and DsRed-PACT (human HeLa cell, gray). Pink dashed line with arrow, centrosome track. Orange dashed lines, cell boundaries. Scale bar, 10 μm. (C) Time-lapse of a pre-abscising human (HeLa) cell expressing centrin-GFP (inverted grays). Green dot, centroid of cell. Orange dashed lines, cell boundaries. Brown line, width of the cytokinetic bridge. Cyan line, distance of the centrosome from the midpoint of the cytokinetic bridge (μm). Scale bar, 10 μm. (D) Model depicting quantification related to centrosome movement towards the cytokinetic bridge. Magenta circle, centrosomes. Black line, cell boundaries. Brown line, width of the cytokinetic bridge. Cyan line, distance of the centrosome from the midpoint of cytokinetic bridge (μm). (E) Percentage of pre-abscising cells (%) that moves the centrosome to at least 2, 4, 6, and 8 μm from the midway point of the cytokinetic bridge.

Figure 2.. Mitotic centrosomes associate with Rab11-endosomes as they reorient towards the cytokinetic bridge.

(A) Time-lapse of pre-abscising human (HeLa) cell expressing DsRed-PACT (inverted grays, top panel; gray in merge, bottom panel) and GFP-FIP3 (cyan in merge, bottom panel). Video 2 Pink arrow, centrosome. Dashed lines, cell boundaries. Scale bar, 10 μm. (B) Time-lapse of a pre-abscising human (Hela) cell expressing Dendra-Rab11. Dendra-Rab11 is photoconverted from 507 nm emission (cyan in merge, bottom panel) to 573 nm emission (inverted grays, top panel; magenta in merge, bottom panel) using 405 nm light within an region of interest over the centrosome (pink arrow) at 0 min. Dashed lines, cell boundaries. Scale bar, 10 μm. (C) Violin plot with median (orange dashed line) and quartiles (black lines) depicting fold change of photoconverted Dendra-Rab11 (573 nm emission) increase at the cytokinetic bridge. Each dot in the plot represents a cell and the different colors depict separate experiments. n = 6 cells and n = 4 experiments. Two-tailed t test, ***P < 0.001. n-values and statistical results detailed in Table S1. (D) A zebrafish embryo pre-abscising cell expressing mCh-Rab11 (gray) fixed and immunolabeled for γ-tubulin (magenta) and DNA (DAPI, blue). Dashed lines, cell boundaries. Scale bar, 10 μm. Insets (d’ and d’’), 2X magnification. (E) Model depicting centrosomes (green) containing Rab11-endosomes (orange) reorienting towards the cytokinetic bridge with associated midbody (purple) during pre-abscission.

Figure S2.. Mitotic centrosomes associate with Rab11 endosomes as they reorient towards the cytokinetic bridge.

(A) Time-lapse of a pre-abscising human (HeLa) cell expressing GFP-FIP3 (16-color LUT). Bottom panels depict three-dimensional surface plot of the intensities portrayed in top panels. Dashed lines, cell boundaries. Scale bar, 5 μm. (A, B) Line graph depicting GFP-FIP3 normalized fluorescence intensity from anaphase exit to late abscission at the polar compartments from (A). Intensities of the brighter pole (pole 1), dark magenta and the less bright pole (pole 2), light pink. (C) Time-lapse of a pre-abscising human (HeLa) cell expressing MannII-mRuby2 (16-color LUT). Bottom panels depict three-dimensional surface plot of the intensities portrayed in top panels. Dashed lines, cell boundaries. Scale bar, 5 μm.

Figure 3.. The oldest mitotic centrosome in pre-abscission moves towards the cytokinetic bridge first and has a more dynamic population of Rab11.

(A, B) Time lapse of a pre-abscising human (HeLa) cell (A) and zebrafish embryo cell at epiboly (B), expressing centrin-GFP (fire LUT). Orange dashed lines, cell boundaries. Scale bar, 10 μm. (C) Percentage of daughter cells with centrosome movement towards the cytokinetic bridge in zebrafish embryos and human (HeLa) cells. n-values on graph, Table S1. (D) Pre-abscising human (HeLa) cell expressing centrin-GFP (fire LUT, magenta in merge) and mCh-Rab11 (gray). (d’) Three-dimensional surface plot of centrin-GFP to note oldest and youngest centrosome. Orange dashed lines, cell boundaries. Scale bar 10 μm. (E) FRAP traces of mCh-Rab11 at the oldest mitotic centrosome (magenta) and youngest mitotic centrosome (green) from three cells depicted. (F) mCh-Rab11 mobile fraction and half-life (T1/2) were calculated as a ratio of oldest centrosome/youngest centrosome and presented as a box and whisker plot with mean (dark line). n = 3. Minimum and maximum values noted by the boxed boundaries. n-values and statistical results detailed in Table S1. (G, H) Expanded pre-abscising human (HeLa) cell expressing centrin-GFP (cyan) and immunostained with Rab11 (magenta) and DNA (DAPI, blue). Dashed lines, cell boundaries. Scale bar, 25 μm. (g’) Magnified insets from (G) depicting Rab11 (gray, magenta in merged) and centrin-GFP (gray, cyan in merged). (H) Example pre-abscising expanded cell centrosome with a model below demonstrating Rab11-endosome area at the mother centriole. (g’, H) M, denotes mother centriole and D, denotes daughter centriole. Scale bar, 2.5 μm. (I) Endosome area at daughter or mother centriole from pre-abscising cells quantified and presented as a violin plot with median (thick line) and quartiles (dotted lines), n = 7 centrosomes, across n = 4 cells. Two-tailed t test, **P < 0.01. n-values and statistical results detailed in Table S1.

Figure S3.. Rab11 GTPase function is associated with centrosome bridge-directed movement.

(A) Equation to calculate distance traveled by centrosome in relation to the cell. This involves recording movement of the centrosome and the cell body as vectors and using vector subtraction to calculate centrosome movement within the reference frame of the cell to accurately calculate distance traveled by the centrosome. (B, C) Time lapse of a pre-abscising human (HeLa) cell (B) and zebrafish embryo cell (C). (A) Inset with associated track (color gradient represents time) of distance traveled by centrosome (cell body centered for each time point) using equation in (A). Scale bar, 10 μm. Black dashed lines, cell boundaries. Color gradient line, time. (D) Western blot immunolabeled for Rab11 and Pericentrin with GAPDH loading control in control cells, Rab11-null cells and Rab11-null cells ectopically expressing mCh-Rab11, -Rab11(S25N), and -Rab11(Q70L). (E, F) Representative images of fixed interphase human (HeLa) cells with a single nucleus or bi-nuclei labeled using DAPI (gray, left; cyan, merge) and Centrin-GFP (magenta) shown (E). Orange dashed lines, cell boundaries. Scale bar, 10 μm. (F) Percentage of binucleate cells calculated for n > 100 cells counted per experiment across n = 3 experiments (F). One-way ANOVA with Dunnett’s multiple comparison with control cells shown, n.s. not significant and **P < 0.01. (G) Total distance traveled by centrosome during pre-abscission measured and depicted as a scatter plot with the median (orange dashed line) and quartiles (dark lines) shown. n > 8 cells per condition across n > 3 experiments. Two-tailed t test, ****P < 0.0001. (H, I, J, K, L) Rab11-null cells ectopically expressing mCh-Rab11 and -Rab11(Q70L) shown (fire, LUT). Region of interest photobleached highlighted in inset (right), pre- and post-FRAP. Scale bar 10 μm (H). mCh-Rab11 (blue dots, lines) and -Rab11(Q70L) (cyan dots, lines) mean fluorescent intensity calculated and presented as a bar graph with mean ± SEM, n > 15 cells per condition. (I) Two-tailed t test, n.s. not significant (I). (J) FRAP trace of mCh-Rab11 and -Rab11(Q70L) shown for n > 15 cells per condition (J). (K, L) mCh-Rab11 and -Rab11(Q70L) mobile fraction (K) and half-life (t1/2, L) calculated and presented as violin plot with median (dark blue line, mCh-Rab11; cyan line, mCh-Rab11(Q70L)) and quartiles (dark blue dotted lines, mCh-Rab11; cyan dotted lines, mCh-Rab11(Q70L)) for n > 15 cells per condition. Two-tailed t test, **P < 0.01 and ****P < 0.0001. n-values and statistical results detailed in Table S1.

Figure 4.. Rab11 GTPase function is associated with centrosome bridge-directed movement.

(A) Control and Rab11-null expressing GFP-FIP3 (cyan, merge) human (HeLa) cells fixed and immunostained for Rab11 (inverted grays, left; magenta, merge). (B) Time-lapse of control and Rab11-null centrin-GFP (inverted grays) pre-abscising human (HeLa) cells. Pink arrows, centrosome. Dashed lines, cell boundaries (A, B). Scale bar, 10 μm Video 3. (C, D, E, F) Rab11-null cells expressing mCh-Rab11, -Rab11 (S25N), or -Rab11(Q70L). (C) Number of centrosomes per pre-abscising cell with bridge directed centrosome movement calculated as both centrosomes (2 centrosomes), only one centrosome (1 centrosome) and neither centrosome (0 centrosomes) moved shown as a violin plot with median (orange dashed line) and quartiles (dark dotted lines). One-way ANOVA, with Dunnett’s multiple comparison to control, ***P < 0.001, **P < 0.01 and n.s. not significant. (D) GFP-FIP3 Rab11-null cells expressing mCh-Rab11, -Rab11(S25N) or -Rab11(Q70L) (inverted grays, left; cyan, merge) fixed and immunostained for Pericentrin (magenta, merge). (E) Centrin-GFP (inverted grays) Rab11-null pre-abscising human (HeLa) cells expressing mCh-Rab11, -Rab11(S25N), or -Rab11(Q70L). Pink arrows, centrosome. Dashed lines, cell boundaries (D, E). Scale bar, 10 μm Video 4. (F) Directional displacement of centrosome towards cytokinetic bridge (scatter plot). Median (orange dashed line) and quartiles (dark lines) shown. One-way ANOVA with Dunnett’s multiple comparison to control, ***P < 0.001 and n.s. not significant. (C, F) n values, statistical results detailed in Table S1.

Figure S4.. Rab11 GTP cycling mediates centrosome protein, Pericentrin, centrosome localization during pre-abscission.

(A, B) Pre-abscising zebrafish embryonic cell (A) or a Rab11-null human (HeLa) cell expressing mCh-Rab11 (gray, A and B top panel; cyan in merge, B) fixed and immunostained for γ-tubulin (magenta, A), Pericentrin (magenta, B), and DNA (DAPI, blue, A). Scale bar, 10 μm. Magnified insets of acentrosomal site (a’, a”, and b”) and centrosome (b’). (C) Time-lapse of GFP-FIP3/Rab11 (inverted grays, left; cyan, merge) centrosome region in a pre-abscising human (HeLa) cell expressing DsRed-PACT (inverted grays, left; magenta, merge). Pink arrows, acentrosomal fragments positive for GFP-FIP3/Rab11/DsRed-PACT. Scale bar, 10 μm. (c’) Outline of tracked particles noted by pink arrow. Orange line, centrosome. (D, E) Human (HeLa) pre-abscising cells expressing GFP-FIP3 (cyan, D) or centrin-GFP (cyan, E) were fixed and immunostained for transferrin receptor (fire LUT, D), and cenexin (fire LUT, E). Magnified insets (3×) shown on right (d’, e’) with associated three-dimensional surface plot of intensity. Scale bar, 5 μm. (F) Human GFP-FIP3 (HeLa) cells late in pre-abscission were fixed and immunostained for Pericentrin (fire LUT). The same examples are shown with levels adjusted on right showing that there is some Pericentrin at one of the daughter centrosomes, but it is significantly decreased. (G) Violin plot with median (orange dashed line) and quartiles (dark lines) depicting normalized Pericentrin intensities at centrosomes for n > 20 centrosomes from a representative experiment. Two-tailed t test, ****P < 0.0001. n-values and statistical results detailed in Table S1.

Figure 5.. Rab11 GTP cycling mediates centrosome protein, Pericentrin, centrosome localization during pre-abscission.

(A) Human (HeLa) pre-abscising cells expressing GFP-FIP3 (cyan) were fixed and immunostained for Pericentrin (magenta, top panel; fire LUT, bottom panel). Magnified insets (3×) shown on right (a’) with associated three-dimensional surface plot of intensity. Scale bar, 5 μm. (B) Box and whisker plot with mean (orange dashed line) depicting ratio of Rab11-null centrosome intensity over control centrosome intensity of cenexin, centrin, Pericentrin, FIP3, and transferrin receptor. Minimum and maximum values noted by boxed boundaries. n > 30 centrosomes per experiment across n = 3 experiments. One-way ANOVA with Dunnett’s multiple comparison to centrin, n.s. not significant, *P < 0.05 and ***P < 0.001. (C) GFP-FIP3 Rab11-null pre-abscising human (HeLa) cells ectopically expressing mCh-Rab11, -Rab11(S25N) or -Rab11(Q70L) (magenta, top panel) were fixed and immunolabeled for acetylated-tubulin (gray, top panel) and Pericentrin (fire LUT, bottom panel). Scale bar, 5 μm. (c’) 3× magnified centrosome inset (left), 3D fluorescent intensity surface plot of inset (right). (D) Scatter plot with median (orange dashed line) and quartiles (dark lines) depicting normalized Pericentrin intensities at centrosomes from pre-abscising human (HeLa) cells. One-way ANOVA with Dunnett’s multiple comparison to control, n.s. not significant and ****P < 0.0001. n > 38 centrosomes across n > 3 experiments. (B, D) n values and statistical results detailed in Table S1.

Figure S5.. Pericentrin and Rab11 endosomes coordinate centrosome movement and number during mitotic exit.

(A) Model depicting generation of pcnt^+/− embryos with labeled centrosomes (centrin-GFP) to test the role of Pericentrin in centrosome reorientation and centrosome number in vivo. (B) Model depicting optogenetic clustering protocol for Rab11 endosomes. In short, embryos are injected with CRY2 and/or CIB1-mCh-Rab11 mRNAs, exposed to blue light 3.3–5 hpf causing a heterointeraction between CRY2 and CIB1, and imaged. Orange inset depicts cells imaged within the embryo. Pink circle, centrosome. (C) Ethidium bromide–stained acrylamide gel electrophoresis showing pcnt^+/+ and pcnt^+/− genotypes with ladder (lane 1), no template (lane 2), DNA extracted from control zebrafish (pcnt^+/+, lane 3) and pcnt^+/− embryos used in experimental analysis in Fig 6 (lane 4, 5, 6). (D) Time-lapse projections depicting a control dividing cell injected with CRY2 and PLK1-mCh (gray). Pink arrow, centrosome. Dashed lines, cell boundaries. Scale bar, 10 μm. (E, F) Representative images of fixed interphase zebrafish embryonic cells with a single nucleus or bi-nuclei labeled using DAPI in control, pcnt^+/−, CIB1-mCh-Rab11 injected, and CRY2 plus CIB1-mCh-Rab11 injected embryos (gray, nuclei outlined with dashed orange line; E, F). Scale bar, 5 μm. (G) Percentage of embryos with cells containing binucleated cells was calculated for n > 30 cells per embryo across n > 3 embryos (G). One-way ANOVA with Dunnett’s multiple comparison with control, n.s. not significant, *P < 0.05 and **P < 0.01. n-values and statistical results detailed in Table S1.

Figure 6.. Pericentrin and Rab11 endosomes coordinate centrosome movement and number during mitotic exit.

(A, B) Time-lapse imaging of control cells (A), pcnt^+/− cells (A), and cells with clustered Rab11 endosomes (B, cyan) from a -5actb2:cent4-GFP(centrin-GFP, gray) embryo. Pink arrows, centrosome. Dashed orange line, cell boundaries. Scale bar, 10 μm. (C, D) Number of centrosomes per pre-abscising cell with bridge directed centrosome movement calculated as both centrosomes (2 centrosomes), only one centrosome (1 centrosome) and neither centrosome (0 centrosomes) moved depicted as a violin plot with median (orange dashed line) and quartiles (dark dotted lines, C), and distance traveled by the centrosome after anaphase (scatter plot with median, orange dashed line, and quartiles, dark lines, D) in control, pcnt^+/−, CRY2 injected, or CRY2 plus CIB1-mCh-Rab11 injected embryos is shown. n > 16 cells across n > 3 embryos. One-way ANOVA with Dunnett’s multiple comparison to control, n.s. not significant, *P < 0.05, **P < 0.01 and ***P < 0.001. (E, F) Interphase centrin-GFP (gray) zebrafish embryos with n = 2 centrosomes or supernumerary (n > 2 centrosomes) in control, pcnt^+/−, CRY2 injected, CRY2 plus CIB1-Rab11 injected embryos (cyan, CIB1-mCh-Rab11 plus CRY2). (E) Orange dashed lines, cell boundaries. Scale bar, 5 μm. (F) Percentage of cells with supernumerary centrosomes (n > 2 centrosomes). n > 30 cells per embryo across n > 3 embryos. One-way ANOVA with Dunnett’s multiple comparison to control, n.s. not significant and ***P < 0.001. (C, D, F) n values and statistical results detailed in Table S1.