EB1 is required for spindle symmetry in mammalian mitosis

Abstract

Most information about the roles of the adenomatous polyposis coli protein (APC) and its binding partner EB1 in mitotic cells has come from siRNA studies. These suggest functions in chromosomal segregation and spindle positioning whose loss might contribute to tumourigenesis in cancers initiated by APC mutation. However, siRNA-based approaches have drawbacks associated with the time taken to achieve significant expression knockdown and the pleiotropic effects of EB1 and APC gene knockdown. Here we describe the effects of microinjecting APC- or EB1- specific monoclonal antibodies and a dominant-negative EB1 protein fragment into mammalian mitotic cells. The phenotypes observed were consistent with the roles proposed for EB1 and APC in chromosomal segregation in previous work. However, EB1 antibody injection also revealed two novel mitotic phenotypes, anaphase-specific cortical blebbing and asymmetric spindle pole movement. The daughters of microinjected cells displayed inequalities in microtubule content, with the greatest differences seen in the products of mitoses that showed the severest asymmetry in spindle pole movement. Daughters that inherited the least mobile pole contained the fewest microtubules, consistent with a role for EB1 in processes that promote equality of astral microtubule function at both poles in a spindle. We propose that these novel phenotypes represent APC-independent roles for EB1 in spindle pole function and the regulation of cortical contractility in the later stages of mitosis. Our work confirms that EB1 and APC have important mitotic roles, the loss of which could contribute to CIN in colorectal tumour cells.

Figure 1. Functional characterisation of antibodies used in this study.

Panel A. The epitope recognised by the EB1-specific monoclonal antibody 1A11 was mapped by co-immunostaining of COS-7 cells transiently transfected with a GFP-tagged EB1 deletion series (Table 1, [17]) and EB1/EB2-GFP chimeric proteins using anti-GFP antibodies (green) and the 1A11 antibody (red). DNA was counterstained using DAPI (blue). The presence of the 1A11 epitope was confirmed by co-localisation of GFP and 1A11 staining. Chimera1 is an EB1/EB2 hybrid protein, while chimera 2 is an EB2/EB1 hybrid. This analysis indicated that the 1A11 epitope lies between aa131–168. Scale bar = 10 µm. Panel B. Specificity of antibody 1A11 is revealed by RNAi experiments. Binding of EB1 is greatly reduced in lysates after EB1 knockdown (lane 3), but not after EB2 (lane 4) and EB3 (lane 5) knockdown. Binding to EB1 is also present in the negative control (scrambled RNA, lane 2) and in normal HeLa cell lysates (lane 1). Similar results are obtained with commercial EB1 antibody clone 5. EB1 knockdown (lane 7) reduced binding of this antibody, whereas knockdown of EB2 (lane 8) or EB3 (lane 7) or with scrambled oligos (lane 6) did not affect EB1 binding. Panel C. EB1 is recognized in lysates from COS7 and NRK-52E cells by 1A11 after Western blotting. Panel D. Effects of 1A11 binding on EB1 interactions with ligands. GST-EB1 bound to glutathione-agarose beads was pre-incubated with 1A11 or non-specific IgG (4 U) before being used in pull-downs from cell extracts. Bound proteins in each pull down were revealed by SDS-PAGE and Western blotting and the effects of antibody pre-incubation compared. Protein levels were quantified using Quantity One and are shown in histograms after background subtraction, corresponding band patterns are shown above the histograms. 1A11 pre-incubation inhibited the binding of GST-EB1 to p150Glued, had little detectable effect on the association with MCAK, and promoted the interactions with CLIP-170 and APC-C3-GFP, an APC C-terminal fragment containing the EB1- and MT-binding regions in APC. Panel E. Effects of the C-APC 9.9 antibody on the EB1 interaction with APC. The anti-APC antibody C-APC9.9 or control IgG 4 U was added to extracts prepared from cells expressing APC-C3-GFP. The ability of GST-EB1 to precipitate APC-C1-GFP from these extracts was then examined. The binding of C-APC9.9 to its epitope in the MT-binding region of APC enhanced the interaction between GST-EB1 and APC-C1-GFP. Protein levels were quantified by Quantity One and presented as histograms after background subtraction, corresponding band patterns are shown above the histogram. Panel F. Endogenous EB1 is not displaced from MT ends in NRK-52E cells microinjected with 1A11. Immunofluorescence images reveal localization of EB1 at MT ends after cells microinjected with 1A11 were fixed and stained with the EB1 antibody clone 5. Fluorescently labelled dextran (red; panel b) indicates a successful microinjection. Immunostaining with clone 5 (green; panel a) reveals EB1 at MT ends. DNA was counterstained using DAPI (blue). Scale bar = 10 µM.

Figure 2. Spindle morphology in failed mitoses.

Representative examples of cells which failed to complete mitosis following injection with either 4 U (panels a–d) or 1A11 (panels e–h). Fluorescently labelled dextran (red; panels b, d, f and h) indicates a successful microinjection. Immunostaining with an anti-tubulin antibody (green; panels a, d, e and h) reveals that these cells do not contain a properly organised mitotic spindle. Condensed DNA is revealed by the addition of DAPI (blue; panels c, d and g, h) Panels d and h show the merged images. Scale bar = 5 µM.

Figure 3. Mitotic progression in microinjected cells.

Stills from representative time-lapse movies of mitotic NRK-52E cells microinjected with the five different monoclonal antibodies used in this study. A: cell microinjected with the non-specific control IgG 4 U. B: cell microinjected with the EB1 specific antibody 1A11. C: cell microinjected with the APC N-terminus specific antibody ALI 12–28, D: cell microinjected with the APC basic domain specific antibody C-APC 9.9. E: cell microinjected with the APC C-terminus specific antibody C-APC 28.9. Scale bar = 10 µM. a-g stills show different stages of mitosis, a- prophase, immediately after microinjection, b- prophase, c- metaphase, d- early anaphase, e- telophase, f–h- telophase/cytokinesis. The arrows indicate the position of the chromosomes. F. The time spent by injected cells in different stages of mitosis is summarised in this graph. Time point 1 (TP1) represents the time taken from nuclear envelope breakdown to anaphase onset. TP2 represents the time taken from anaphase onset to the start of telophase. TP3 corresponds to the time taken from the start of telophase until the appearance of a phase-dense midbody during cytokinesis. No significant differences in the duration of these time points was seen in cells injected with the different antibodies until TP3, which cells injected with 1A11 (p<0.05) and ALI 12–28 (p<0.005) took significantly longer to transit. The graph shows the mean time in min +/− SEM for each condition. Significant results are indicated with an asterisk. (4 U, n = 12; 1A11, n = 9; ALI12–28, n = 9; C-APC 9.9, n = 14; C-APC 28.9, n = 11).

Figure 4. Microinjection with EB1 or APC specific antibodies affects chromosomal congression at the metaphase plate.

A: Movies of microinjected cells were examined and the last frame before anaphase onset identified. The distance separating the widest two points of the metaphase plate was then measured as indicated. In the representative examples shown it can be seen that the metaphase plate was more tightly compacted in cells injected with control IgG (panel a) than in cells microinjected with 1A11 (panel b). Scale bar = 5 µM. The red traces indicate the outline of the metaphase plates. B: Summary of the metaphase plate thickness data. The metaphase plate was significantly wider in cells microinjected with APC or EB1 specific antibodies when compared to cells injected with the non-specific antibody 4 U. Mean distances +/− SEM are shown. Significant results are shown with an asterisk. P values were for 1A11, p<0.005; ALI 12–28, p<0.005; C-APC 9.9, p<0.05 and C-APC 28, p<0.005. (4 U, n = 11; 1A11, n = 14; ALI 12–28, n = 11, C-APC 9.9, n = 12; C-APC 28.9, n = 7).

Figure 5. Effects of antibody injection on mitotic spindle positioning.

Following antibody microinjection in prophase, spindle alignment and placement was assessed. A: Still image of a cell microinjected with 1A11 at metaphase displaying spindle displacement and misalignment, (a) the black asterix indicates the geometrical centre of the cell and the yellow asterix the centre of the metaphase plate; (b) the white line indicates the longitudinal axis of the cell and the black line the axis of the spindle. B: Spindle alignment in microinjected cells at TP1 (onset of anaphase) and TP2 (onset of telophase). At TP1 the greatest amount of spindle misalignment was observed among cells microinjected with 1A11 or C-APC 9.9 and these cells were also less likely to be correctly aligned at TP2. However, these differences did not achieve statistical significance. (4 U, n = 14; 1A11, n = 13, ALI 12–28, n = 11, C-APC 9.9, n = 9, C-APC 28.9, n = 11). C: Spindle placement within the mitotic cell at TP1 was unaffected by APC antibody microinjection, although spindles were displaced further from the geometric centre of the cell in cells injected with the EB1 specific antibody 1A11. However, this failed to reach statistical significance (p = 0.06). Mean spindle displacement +/− SEM is shown. (4 U, n = 11; 1A11, n = 14; ALI 12–28, n = 11, C-APC 9.9, n = 12; C-APC 28.9, n = 7). D: Movie stills taken from a recording of a representative cell microinjected with 1A11 and fixed at anaphase. Note that at metaphase (51:00 min) the spindle is misplaced and misaligned. Scale bar = 10 µm. For movie of these stills see Movie S6. The red outline shows cell shape, the white line indicates the longitudinal axis of the cell and the black line the axis of the spindle.

Figure 6. EB1 antibody injection induces anaphase-specific cortical blebbing.

A: Stills from a recording of a representative cell microinjected with 1A11, showing the appearance of cortical blebbing during anaphase (arrows). Scale bar = 10 µM. B: The percentage of cells displaying cortical blebbing is first shown as a percentage of the total number of microinjected cells examined (including those that failed to complete mitosis) and then as a percentage of microinjected cells that successfully completed mitosis. In both cases a significantly higher proportion of cells microinjected with antibody 1A11 displayed blebbing (p<0.005 and p<0.05). 1A11-induced blebbing was only seen during anaphase whereas a lower level of blebbing was noted throughout mitosis in cells injected with anti-APC antibodies (see Figure 2 panels D and E for examples of blebbing in metaphase cells injected with anti-APC antibodies). Blebbing was not observed in mitotic cells injected with the non-specific IgG 4 U (zero entry in first column). Significant results are indicated by an asterisk. (4 U, n = 16 for total number of microinjected cells, n = 12 for cells with completed mitosis; 1A11, n = 19 and 11; ALI 12–28 n = 16 and 9; C-APC 9.9, n = 18 and 16; C-APC 28.9, n = 15 and 11).

Figure 7. Microinjection of 1A11 induces asymmetric spindle pole movement.

A: Stills (a–f) taken from a recording of a representative cell microinjected with 1A11. As shown by the tracks, one set of chromosomes (the black spot closest to the top of the frame) remains relatively static within the cell during anaphase B and telophase while the other moves. The black and red tracks indicate chromosomal movement. Scale bar = 10 µM. B: Velocities were obtained for anaphase chromosomal movement in microinjected cells and used to derive a ratio that reflected any asymmetry in the observed movements, with values close to 1 representing symmetrical chromosomal separation. A plot showing the range of velocity ratios obtained is shown. Only cells injected with the EB1 specific antibody 1A11 displayed an asymmetry in movement that generated a ratio of less than 0.5 (column 2). (4 U, n = 14; 1A11, n = 14; ALI 12–28, n = 10, C-APC 9.9, n = 12; C-APC 28.9, n = 9). Black line indicates mean value.

Figure 8. Asymmetrical microtubule content in daughters of cells injected with 1A11.

Microinjected cells were allowed to complete mitosis then were fixed and immunostained to reveal the MT cytoskeleton. A: Images obtained from daughter cells after microinjection of mothers with 4 U and 1A11. Daughters of cells injected with the control antibody 4 U were equally sized and displayed an equal content of microtubules (a–c), whereas the daughters of 1A11 injected cells exhibited more uneven microtubule contents (d–i) and in some cases were of unequal sizes (d–f). Scale bar = 5 µM. B. The fluorescence content of daughter cells immunostained for tubulin was measured and used to derive a ratio that reflected any asymmetry in the MT content of the cells, with values close to 1 representing essentially equal MT contents. A plot showing the range of ratio values obtained is shown. The daughters of cells injected with 1A11 were more likely to exhibit asymmetry in MT content than cells injected with other antibodies. (4 U, n = 6; 1A11, n = 9; ALI 12–28, n = 5; C-APC 9.9, n = 6; C-APC 28.9, n = 7). C: A plot comparing the spindle pole movement and microtubule intensity ratios obtained for 4 U injected cells, n = 5 and 1A11 injected cells, n = 8). Mitoses where severely asymmetrical pole movement was seen also generated daughters with the most asymmetrical MT contents, with daughters inheriting the least mobile pole containing fewest MTs. Here, the same coloured shape indicates results for the same cell within either the 4u- microinjected group of cells or the 1A1-microinjected group of cells.

Figure 9. Microinjection of an anti-EB1 antibody induces tilted spindles.

A. Immunostaining images of 4 U microinjected cells (a) and 1A11 (b,c) fixed and stained at anaphase. The images for the 1A11 injected cell had to be taken separately as the spindle poles were not in the same focal plane. Arrows indicate the two spindle poles, which were not in the same focal plane. Tubulin in green, combined anti-EB1 and dextran staining in red, DAPI in blue. Scale bar = 5 µM. B. Immunostaining image of 1A11 microinjected cells at metaphase. Arrow indicates a microinjected cell whose spindle poles are not in the same focal plane, revealing the presence of a spindle tilted in relation to the substrate. Tubulin in green, injected 1A11 and fluorescent dextran staining in red, DAPI in blue. Scale bar = 10 µM.

Figure 10. Characterisation of antibody 1A11 binding to EB1 in mitotic NRK-52E cells.

A. Binding pattern of the EB1 specific antibody 1A11 in mitotic NRK-52E cells. The antibody binds to the spindle poles in fixed and immunostained cells (a–c) and microinjected, fixed and stained cells (d–f). Tubulin green, EB1 red, DAPI blue. Scale bar = 5 µM. B. EB1 staining patterns of spindle poles at prophase appear to be uneven as indicated by immunofluorescence of fixed and stained NRK-52E cells. EB1 green (panel a), centrin red (panel b), DAPI blue. Scale bar = 5 µM. C. EB1 is more unevenly distributed on centrosomes in prophase than in the later stages of mitosis, especially when compared to anaphase, in both the presence and absence of nocodazole (significant results are indicated with an asterisk, p<0.05).

Figure 11. Microinjection of an inhibitory EB1 fragment into NRK-52E cells recapitulates phenotypes observed in EB1 antibody injected cells.

Mitotic NRK-52E cells were microinjected with the dominant-negative EB1 fragment GST-EB1-C84 or with GST alone. A) Injection of low levels of GST-EB1-C84 inhibited progression through mitosis. B) Cytokinesis was significantly delayed in mitotic cells injected with GST-EB1-C84 that completed mitosis (p = 0.01). C) Chromosomal congression to the metaphase plate was significantly inhibited in cells microinjected with GST-EB1-C84. D) The majority of cells microinjected with GST-EB1-C84 had misaligned spindles at TP1, most of which had not corrected by the time cells reached telophase. E) Some cells microinjected with GST-EB1-C84 showed severe asymmetry in spindle movement, as shown by a ratio value of <0.5. Different colours and shapes represent individual cells. (F) Cells injected with GST-EB1-C84 tended to possess spindles that were placed further from the cell centre, though this trend did not reach statistical significance. (G) A higher proportion of cells injected with GST-EB1-C84 exhibited cortical blebbing, though it was not mitotic stage specific.

Figure 12. Microinjection of mitotic NRK-52E cells with GST or GST-EB1-C84 fusion protein.

Stills from representative time-lapse movies of mitotic NRK-52E cells after microinjection with GST alone (A, a–h) or GST-EB1-C84 (B, a–h). Black arrows indicate the position of the chromosomes. Note cortical blebbing and the formation of an ectopic cleavage furrow, which is resolved by the time the GST-EB1-C84 injected cell reaches cytokinesis (B, red arrow) C: The cells shown in A and B were fixed and immunostained with an antibody specific for alpha-tubulin (green) and counterstained using DAPI (blue). Injected cells were identified by the presence of fluorescent dextran (red). The cells shown in a and b are daughter cells of the cell injected with GST, while the cells in c and d are daughter cells of the cell injected with GST-EB1-C84. Note the formation of uneven daughter cells after microinjection with the GST-EB1-C84 protein. Merged images are shown in b and d. Scale bars = 10 and 5 µm.

Figure 13. Diagram showing the possible interactions of EB1 with main interactants and the effect of antibody binding.

Blocking of EB1 with 1A11 enhances some interactions (CLIP-170, APC), reduces interactions with p150Glued and has no effect on other interactions, for example with MCAK (shifted balance after preincubation with specific antibody).

Figure 14. Models for the possible functional effects of antibody 1A11 in microinjected cells.

(A) In uninjected cells EB1 (orange triangle) is initially localised to the mother spindle pole (blue rectangle) and is recruited to the daughter (black rectangle) during prophase, giving both poles functional equivalency in MT nucleation and/or anchoring during and after mitosis. (B) In microinjected cells the antibody binds preferentially to EB1 located at the mother spindle pole early in prophase and antagonises its function. The function of EB1 newly recruited to the daughter pole is not affected, leading to proper nucleation and/or anchoring of microtubules at the daughter but not the mother pole. The uneven distribution of functional EB1 at the poles leads to asymmetries in mitotic pole movement and MT content in daughter cells. (C) EB1 at the mother pole is associated with a binding partner (green circle) important for its function at this site. Antibody binding does not antagonise the function of pre-existing complexes, but it inhibits the functional associations of EB1 newly recruited to the daughter pole. This results in normal nucleation and/or anchoring of microtubules at the mother but not the daughter pole, again leading to asymmetries in mitotic pole movement and MT content in daughter cells.