Building a spindle of the correct length in human cells requires the interaction between TPX2 and Aurora A

Abstract

To assemble mitotic spindles, cells nucleate microtubules from a variety of sources including chromosomes and centrosomes. We know little about how the regulation of microtubule nucleation contributes to spindle bipolarity and spindle size. The Aurora A kinase activator TPX2 is required for microtubule nucleation from chromosomes as well as for spindle bipolarity. We use bacterial artificial chromosome-based recombineering to introduce point mutants that block the interaction between TPX2 and Aurora A into human cells. TPX2 mutants have very short spindles but, surprisingly, are still bipolar and segregate chromosomes. Examination of microtubule nucleation during spindle assembly shows that microtubules fail to nucleate from chromosomes. Thus, chromosome nucleation is not essential for bipolarity during human cell mitosis when centrosomes are present. Rather, chromosome nucleation is involved in spindle pole separation and setting spindle length. A second Aurora A-independent function of TPX2 is required to bipolarize spindles.

Figure 1.

A BAC transgene containing mTPX2-LAP rescues the phenotype of TPX2-RNAi. (a) Western blot of cell lines with anti-mTPX2 antibody, which recognizes both human TPX2 and mTPX2-LAP, and anti–α-tubulin, after TPX2 or CON RNAi. hTPX2 is efficiently depleted in all lines after TPX2 RNAi, whereas mTPX2-LAP transgenes remain. The asterisk represents the position of the TPX2^ΔN protein, which runs faster than the full length, as expected. (b) Immunofluorescence of U2OS cells after CON or TPX2 RNAi and TPX2^WT cells after TPX2 RNAi, stained for α-tubulin (green), Cep135 (red), and DNA (blue). U2OS cells without a rescuing transgene show a characteristic phenotype after hTPX2 depletion of collapsed poles and lack of a bipolar spindle, whereas TPX2^WT cells containing mTPX2-LAP after hTPX2 depletion have normal spindle morphology. Bar,10 μm. (c) Quantification of the TPX2 depletion phenotype and mTPX2-LAP rescue displayed in b. The percent of mitotic cells (n = >1,000 cells per experiment; three to five experiments for each condition) and percent bipolar spindles (n > 50) mitotic cells per experiment; three to five experiments per condition) after CON or TPX2 RNAi in U2OS or TPX2^WT cells were determined. Error bars represent standard deviation. The mTPX2-LAP construct is able to rescue depletion of endogenous TPX2.

Figure 2.

TPX2 mutants abolish the in vivo interaction between TPX2 and Aurora A as well as Aurora A localization to spindles. (a) Western blots of cell extracts immunoprecipitated to detect interaction of TPX2 with Aurora A. U2OS (untagged), mTPX2^WT, mTPX2^AAA, and mTPX2^ΔN cells were arrested in mitosis with nocodazole, and protein extracts were immunoprecipitated with anti-GFP antibody. Input and immunoprecipitated fractions were run by SDS-PAGE and blotted with either anti-mTPX2 or anti–Aurora A antibody. Anti-GFP antibody pulls down mTPX2^WT bound to Aurora A, whereas nothing is immunoprecipitated in the untagged cells. Immunoprecipitating mTPX2^WT does not pull down endogenous TPX2. In mTPX2^AAA and mTPX2^ΔN cells, the mTPX2 mutant transgene is pulled down, but Aurora A is not. The single asterisk represents the position of the TPX2^ΔN protein. The double asterisk represents a degradation product of the mTPX2 protein. (b) Immunofluorescence analysis of cell lines after CON or TPX2 RNAi stained for α-tubulin (green), DNA (blue), mTPX2-LAP (anti-GFP; insets), and Aurora A. Aurora A is localized to spindles and centrosomes in U2OS cells (top row). After hTPX2 RNAi, Aurora A is absent from spindles but still on centrosomes (second row). A mouse TPX2-LAP transgene restores Aurora A localization to spindles after hTPX2 RNAi (third row). The same mouse transgene with point mutations introduced to abolish the TPX2–Aurora A interaction no longer is able to recruit Aurora A to spindles after hTPX2 RNAi (bottom row). The pericentrosomal Aurora A localization shown in the mutant similar to that of the TPX2 depletion was observed in images of 12 of 14 cells. In 2 of 14 images, the Aurora A localization in the mutant reflected a more centrosomal localization. Bar, 10 μm.

Figure 3.

TPX2 localizes to kinetochore fibers independent of interaction with Aurora A. Immunofluorescence analysis of TPX2 localization. (a) U2OS cells in metaphase (top row) and anaphase (bottom row) stained for DNA, CREST, hTPX2 (green), and α-tubulin (black and white). Insets are enlarged to show TPX2 localization to microtubules that end at kinetochores (kinetochore fibers). (b) mTPX2^WT cells after hTPX2(RNAi) with same staining as in a, except with mTPX2 instead of hTPX2 to recognize the transgene. (c) mTPX2^AAA cells after hTPX2(RNAi) with same staining as in b. mTPX2 mutated to abolish interaction with Aurora A is still able to localize to kinetochore-fibers. Bar, 10 μm.

Figure 4.

The TPX2–Aurora A interaction is required for spindle pole separation and spindle length establishment. (a) Immunofluorescence of mTPX2^WT, mTPX2^AAA, and mTPX2^ΔN mitotic cells after hTPX2(RNAi), stained for CREST, Cep135, DNA, and α-tubulin. Centrosomes are collapsed to chromatin when the TPX2–Aurora A interaction is abolished. Bar, 10 μm. (b) Centrosome-kinetochore distances are shorter in mTPX2^AAA and mTPX2^ΔN mutants. The centrosome-kinetochore distance from cells stained as in a was determined by measuring individual centrosome-kinetochore distances from metaphase cells in three dimensions (see Materials and methods). For each condition, measurements were taken from four independent cells (eight centrosomes). Red dots show the data points measured and the bars show the means. (c) In TPX2 mutants unable to bind or activate Aurora A, spindle poles collapse immediately after NEBD and only slightly elongate to form shorter metaphase spindles. Distances between spindle poles were measured in three dimensions from live-cell image stacks taken at 1-min intervals of mTPX2^WT, mTPX2^AAA, and mTPX2^ΔN (mTPX2-GFP) cells after hTPX2(RNAi). Thick lines represent means and thin lines represent individual videos. (d) Representative still images from time-lapse recordings quantified in b.

Figure 5.

TPX2 mutants have a higher percentage of defective divisions, delayed mitoses, and increased metaphase BUB1 staining. (a) Percentage of defective mitoses in mTPX2^WT, mTPX2^AAA, and mTPX2^ΔN cells after TPX2 or CON RNAi, quantified by observing time-lapse videos of GFP fluorescence. Actual numbers of cell divisions observed are shown below the histogram. (b) mTPX2^AAA and mTPX2^ΔN cells after TPX2 RNAi show longer delays in mitosis and a higher percentage of cells exhibiting delays. Lengths of individual mitosis are plotted for mTPX2^WT, mTPX2^AAA, and mTPX2^ΔN cells after TPX2 or CON RNAi. Percentages shown are the fraction of cells showing lengths of mitoses greater than 120 min. (c) Immunofluorescence of mTPX2^WT and mTPX2^ΔN prometaphase and metaphase cells after TPX2 RNAi, stained for BUB1 or MAD2. Insets show DNA (DAPI) from the same cell at lower magnification. In mTPX2^WT cells, MAD2 and BUB1 foci are enriched in prophase but not metaphase. In mTPX2^ΔN cells, MAD2 enrichment is gone in metaphase cells, whereas some Bub1-positive foci remain. Bars, 10 μm.

Figure 6.

Depletion of TPX2 from cells results in a spindle defect phenotype distinct from TPX2 mutants or Aurora A depletion. (a) Distances between spindle poles were measured in three dimensions from live-cell image stacks taken at 3-min intervals of U2OS cells stably transfected with α-tubulin–mCherry after treatment with CON, TPX2, or AurA siRNA. Thick lines represent means and thin lines represent individual videos. The dashed line represents data from Fig. 4 for reference. (b) Representative still images from time-lapse recordings quantified in a. The cells shown transfected with CON or AurA siRNAs completed mitosis and divided, whereas the cell depleted of TPX2 arrested in mitosis for several hours and displayed fragmented spindle poles. Bar, 10 μm.

Figure 7.

The TPX2–Aurora A interaction is required for cold-stable microtubule stability and organization in prometaphase. mTPX2^WT, mTPX2^AAA, and mTPX2^ΔN cells were transfected with TPX2 or CON siRNA. 48 h after transfection, cells were treated on ice for 5 min and fixed. Fluorescence images are shown of cells stained with anti–α-tubulin (microtubules) and DAPI (DNA) showing cold-stable microtubule populations. Bar, 10 μm.

Figure 8.

The TPX2–Aurora A interaction is required for chromatin/kinetochore-mediated microtubule nucleation. U2OS, mTPX2^WT, mTPX2^AAA, and mTPX2^ΔN cells transfected with TPX2 or CON siRNA were treated on ice to completely depolymerize microtubules, transferred to 37°C for 90 s to allow repolymerization, and immediately fixed. Immunofluorescence images are shown of cells stained with α-tubulin and CREST (top rows) and Cep135 and DNA (bottom rows). In addition to aster microtubule polymerization associated with centrosomes, microtubule polymerization broadly associated with chromatin and often specifically associated with kinetochores was evident in all cell lines after CON siRNA and in mTPX2^WT cells after TPX2 RNAi. In U2OS, mTPX2^AAA, and mTPX2^ΔN cells after TPX2 RNAi, microtubule repolymerization was no longer seen associated with chromatin or kinetochores, but centrosome-associated microtubule aster polymerization was still evident. Bar, 10 μm.