Larson ME and Bement WM (2017), Automated mitotic spindle tracking suggests a l...

XB-ART-53016

Mol Biol Cell 2017 Mar 15;286:746-759. doi: 10.1091/mbc.E16-06-0355.

Show Gene links Show Anatomy links

Automated mitotic spindle tracking suggests a link between spindle dynamics, spindle orientation, and anaphase onset in epithelial cells.

Larson ME, Bement WM.

???displayArticle.abstract???
Proper spindle positioning at anaphase onset is essential for normal tissue organization and function. Here we develop automated spindle-tracking software and apply it to characterize mitotic spindle dynamics in the Xenopus laevis embryonic epithelium. We find that metaphase spindles first undergo a sustained rotation that brings them on-axis with their final orientation. This sustained rotation is followed by a set of striking stereotyped rotational oscillations that bring the spindle into near contact with the cortex and then move it rapidly away from the cortex. These oscillations begin to subside soon before anaphase onset. Metrics extracted from the automatically tracked spindles indicate that final spindle position is determined largely by cell morphology and that spindles consistently center themselves in the XY-plane before anaphase onset. Finally, analysis of the relationship between spindle oscillations and spindle position relative to the cortex reveals an association between cortical contact and anaphase onset. We conclude that metaphase spindles in epithelia engage in a stereotyped "dance," that this dance culminates in proper spindle positioning and orientation, and that completion of the dance is linked to anaphase onset.

???displayArticle.pubmedLink??? 28100633
???displayArticle.pmcLink??? PMC5349782
???displayArticle.link??? Mol Biol Cell
???displayArticle.grants??? [+]

Species referenced: Xenopus laevis
Genes referenced: ddx59 h2bc21 neb tub

???attribute.lit??? ???displayArticles.show???

	FIGURE 1:. Tubulin and histone highlight mitotic landmarks. (A) eGFPâ€“Î±-tubulin (GFP-Tub) highlights the highly dynamic mitotic spindle in the X. laevis embryo. (B) X. laevis gastrula animal caps contain a field of asynchronous epithelial cells, visualized with mCherry-histone H2B (mChe-H2B; Bâ€²) and GFP-Tub (Bâ€²â€²). (C) Mitotic temporal landmarks are apparent in cells expressing mChe-H2B and GFP-Tub, including NEB (frames 1 and 2), formation of the metaphase plate (frame 3), and segregation of chromosomes in anaphase (frame 4). The line in frame 4 through the spindle poles at anaphase onset was used to generate a kymograph (D), highlighting NEB (arrowhead), anaphase onset (arrow), and spindle movements in preanaphase period.
	FIGURE 2:. CAAX provides a bright cell-edge marker. (A) Comparison of cortical tubulin and mTagBFP-CAAX (BFP-CAAX, a membrane marker) in interphase cell with accompanying line scan (B) reveals both tubulin and CAAX signal at the cell edge. (C) Comparison of tubulin and membrane probes in mitotic cell with accompanying line scan (D) demonstrates loss of cortical tubulin in mitosis. Cells were labeled through expression of GFP-Tub, mChe-H2B, and BFP-CAAX.
	FIGURE 3:. Mitotic spindles display stereotyped alignment and oscillatory rotational dynamics. The angle between centrosomes/spindle poles and the long axis of the cell at anaphase was measured in 28 cells at (A) NEB, (B) metaphase, and (C) anaphase in cells expressing BFP-CAAX, GFP-Tub, and mChe-H2B. (D) Mitotic spindles display dynamic movement, featuring preanaphase oscillations when spindle orientation is plotted (E). (F) Average spindle orientation over time for 28 manually analyzed spindles. Time 0 indicates approximate time of metaphase plate formation. (G) Twenty-eight spindles all demonstrate rotational oscillations before anaphase onset (time 0). Spindles grouped by net rotation between NEB and anaphase onset show no significant difference in (H) mitotic duration, (I) total oscillations, (J) maximum amplitude of oscillation, or (K) periodicity of oscillation. (L) Power spectral density of corresponding spindle traces from G, indicating the oscillatory frequency of each spindle. (M) Average of PSDs from L, indicating a dominant oscillatory frequency of 0.01562 Hz (period of 64 s).
	FIGURE 4:. Automated analysis detects spindle rotational oscillations. (A) A time series of mitosis in cells expressing mChe-H2B, GFP-Tub, and BFP-CAAX (left), with accompanying automatically recognized regions (middle) and the convex hull of those regions (right), output from the Spindlometer. (B) Manually and (C) automatically calculated plots of spindle rotation demonstrate rotational oscillations before anaphase onset.
	FIGURE 5:. Analysis of oscillations reveals biphasic movement. (A) Average of 100 rotational trajectories reveals net rotation through mitosis, with slight apparent oscillation and decrease in net rotation toward anaphase onset. (B) Rotational traces of two spindles, one that is prealigned with the anaphase axis and one that is not. (C) Average normalized RMS of spindle rotational acceleration over 2-min intervals (24 frames) of 100 spindles during metaphase. This metric will be 0 if the spindle is still or has a constant rotational velocity and increases with greater changes in velocity over the 2-min intervals. RMS is low at early time points and steadily increases to a plateau, indicating that rotational velocity is more constant in early metaphase, consistent with a rotational but nonoscillatory regime, followed by a period of more tumultuous acceleration, indicative of oscillation. Inset, sample RMS traces corresponding to the rotational traces in B.
	FIGURE 6:. Comparison of cell and anaphase spindle metrics. One hundred cells expressing mChe-H2B, GFP-Tub, and BFP-CAAX were analyzed with the Spindlometer. (A) Comparison of angle between anaphase spindle orientation and NEB cell orientation and cell eccentricity reveals decreased angle as cell eccentricity increases (linear regression fit, slope = âˆ’59.6, p < 0.0001). (B) Average cell eccentricity decreased through mitosis (p < 0.0001, paired t test). (C) Angle between spindle and cell orientation is independent of change in cell eccentricity through mitosis (p = 0.24). (D) Comparison of distance between spindle centroid and cell centroid during metaphase and at anaphase onset reveals a significant decrease of displacement (increased spindle centering) at anaphase onset (p = 0.0004, paired t test).
	FIGURE 7:. Analysis of spindle pole movements. (A) Averaged power spectral density of 100 automatically calculated spindle rotational trajectories in the 150 s before anaphase onset reveals a similar maximum to that seen in manually analyzed spindles. (B) Mean square displacement of 200 metaphase spindle pole locations reveals an initial supralinear (directed) behavior and later infralinear (restricted) behavior. Black line represents mean MSD; gray areas indicates SD. (C) Diagram of spindle-based moving reference frame used to highlight spindle dynamics. Cross-correlation of radial velocity components (D) demonstrates correlated movement at zero time lag, as well as outlying peaks (−50 and 50 s) indicative of oscillation.
	FIGURE 8:. Measurement of spindle–cortex interactions. For 100 spindles analyzed with the Spindlometer, spindle pole velocity was broken down into components parallel to (radial) or perpendicular to (axial) the nearest point on the cell boundary. (A) Linear fits to positive axial velocity, negative axial velocity, and nondirectional radial velocity against distance to cell boundary. Radial velocity is significantly greater than either axial component at the cell edge (p < 0.001). Negative axial velocity is significantly greater than positive axial velocity at the cell edge (p < 0.001). (B) Sample trace of distance between spindle pole and cell boundary. Triangles denote troughs where distance drops to <7 µm. C) Average distance of 200 spindle poles to cell boundary during metaphase. Time 0, anaphase onset. Error bars, SE. (D) For 200 metaphase spindle poles, histogram indicating the 2% of time points at which each pole was closest to the cortex, indicating a general increase in cortical contact through metaphase. (E) Timing of cortical contacts for 100 spindles during metaphase. A contact is constituted by proximity within 7 µm of the cell boundary followed by withdrawal by at least 2 µm. (F) For 200 spindle poles, average distance between the spindle pole and points on the cell cortex that fall along the anaphase axis. Time 0, anaphase onset. Error bars, SE. (G) Timing of on-target cortical contacts at the points on the cortex along the anaphase axis. Contact is constituted by proximity within 7 µm, followed by withdrawal of at least 3 µm. (H) Timing of off-target (unaligned) cortical contacts. Cortical contacts as defined in D were sorted to include only those that occurred at least 9.3 µm from the target point defined in G.
	FIGURE 1:. Tubulin and histone highlight mitotic landmarks. (A) eGFPâ€“Î±-tubulin (GFP-Tub) highlights the highly dynamic mitotic spindle in the X. laevis embryo. (B) X. laevis gastrula animal caps contain a field of asynchronous epithelial cells, visualized with mCherry-histone H2B (mChe-H2B; Bâ€²) and GFP-Tub (Bâ€²â€²). (C) Mitotic temporal landmarks are apparent in cells expressing mChe-H2B and GFP-Tub, including NEB (frames 1 and 2), formation of the metaphase plate (frame 3), and segregation of chromosomes in anaphase (frame 4). The line in frame 4 through the spindle poles at anaphase onset was used to generate a kymograph (D), highlighting NEB (arrowhead), anaphase onset (arrow), and spindle movements in preanaphase period.
	FIGURE 2:. CAAX provides a bright cell-edge marker. (A) Comparison of cortical tubulin and mTagBFP-CAAX (BFP-CAAX, a membrane marker) in interphase cell with accompanying line scan (B) reveals both tubulin and CAAX signal at the cell edge. (C) Comparison of tubulin and membrane probes in mitotic cell with accompanying line scan (D) demonstrates loss of cortical tubulin in mitosis. Cells were labeled through expression of GFP-Tub, mChe-H2B, and BFP-CAAX.
	FIGURE 3:. Mitotic spindles display stereotyped alignment and oscillatory rotational dynamics. The angle between centrosomes/spindle poles and the long axis of the cell at anaphase was measured in 28 cells at (A) NEB, (B) metaphase, and (C) anaphase in cells expressing BFP-CAAX, GFP-Tub, and mChe-H2B. (D) Mitotic spindles display dynamic movement, featuring preanaphase oscillations when spindle orientation is plotted (E). (F) Average spindle orientation over time for 28 manually analyzed spindles. Time 0 indicates approximate time of metaphase plate formation. (G) Twenty-eight spindles all demonstrate rotational oscillations before anaphase onset (time 0). Spindles grouped by net rotation between NEB and anaphase onset show no significant difference in (H) mitotic duration, (I) total oscillations, (J) maximum amplitude of oscillation, or (K) periodicity of oscillation. (L) Power spectral density of corresponding spindle traces from G, indicating the oscillatory frequency of each spindle. (M) Average of PSDs from L, indicating a dominant oscillatory frequency of 0.01562 Hz (period of 64 s).
	FIGURE 4:. Automated analysis detects spindle rotational oscillations. (A) A time series of mitosis in cells expressing mChe-H2B, GFP-Tub, and BFP-CAAX (left), with accompanying automatically recognized regions (middle) and the convex hull of those regions (right), output from the Spindlometer. (B) Manually and (C) automatically calculated plots of spindle rotation demonstrate rotational oscillations before anaphase onset.
	FIGURE 6:. Comparison of cell and anaphase spindle metrics. One hundred cells expressing mChe-H2B, GFP-Tub, and BFP-CAAX were analyzed with the Spindlometer. (A) Comparison of angle between anaphase spindle orientation and NEB cell orientation and cell eccentricity reveals decreased angle as cell eccentricity increases (linear regression fit, slope = âˆ’59.6, p < 0.0001). (B) Average cell eccentricity decreased through mitosis (p < 0.0001, paired t test). (C) Angle between spindle and cell orientation is independent of change in cell eccentricity through mitosis (p = 0.24). (D) Comparison of distance between spindle centroid and cell centroid during metaphase and at anaphase onset reveals a significant decrease of displacement (increased spindle centering) at anaphase onset (p = 0.0004, paired t test).

References [+] :

Adames, Microtubule interactions with the cell cortex causing nuclear movements in Saccharomyces cerevisiae. 2000, Pubmed