Myosin II promotes the anisotropic loss of the apical domain during Drosophila neuroblast ingression

Abstract

Epithelial-mesenchymal transitions play key roles in development and cancer and entail the loss of epithelial polarity and cell adhesion. In this study, we use quantitative live imaging of ingressing neuroblasts (NBs) in Drosophila melanogaster embryos to assess apical domain loss and junctional disassembly. Ingression is independent of the Snail family of transcriptional repressors and down-regulation of Drosophila E-cadherin (DEcad) transcription. Instead, the posttranscriptionally regulated decrease in DEcad coincides with the reduction of cell contact length and depends on tension anisotropy between NBs and their neighbors. A major driver of apical constriction and junctional disassembly are periodic pulses of junctional and medial myosin II that result in progressively stronger cortical contractions during ingression. Effective contractions require the molecular coupling between myosin and junctions and apical relaxation of neighboring cells. Moreover, planar polarization of myosin leads to the loss of anterior-posterior junctions before the loss of dorsal-ventral junctions. We conclude that planar-polarized dynamic actomyosin networks drive apical constriction and the anisotropic loss of cell contacts during NB ingression.

Figure 1.

NBs ingress via an anisotropic loss of apical cell–cell junctions. (A) An NB undergoes apical constriction and disassembles its AJs while moving out of the epithelium. (B–D) Stills of a time-lapse video of the ectoderm during stage 8 (membrane marker Spider::GFP). Numbered cells are medial (red, M), intermediate (blue, I), and lateral (yellow, L) NBs, located 1–3, 4, and 7–9 cell rows away from the VM, respectively. White arrowheads indicate the position of NBs that have lost their apical membrane. Insets show the corresponding stage of embryonic axis elongation (black arrowheads, posterior end). Anterior, left; dorsal, up. (E) Stills of a time-lapse video of an NB (apical domain, yellow; ubi-DEcad::GFP). Arrowheads point to new horizontal junctions forming between DV NICs. (F) Watershed segmentation (green outline) of the ingressing NB in E. Bars: (B–D) 15 µm; (E and F) 5 µm. (G) Mean loss of apical cell area during ingression. Time 0 (dashed line) indicates the onset of ingression. (H) Developmental timing of completion of ingression by NB row. Medial NBs complete ingression first, followed by the intermediate and lateral rows. Time 0 corresponds to the onset of cytokinesis in VM cells. Box plot with IQRs and minimum/maximums displayed. *, P = 0.037; ****, P = 0.0001 (two-tailed t test). (I) DV and AP apical cell length during ingression (Spider::GFP, n values as in G). The apical surface shrinks along the DV axis 1.6× faster than along the AP axis. Time 0 is the onset of ingression. Data presented are means ± SEM.

Figure 2.

Sna family proteins and DEcad transcriptional down-regulation are not essential for NB ingression. (A–C) Embryos expressing Sna::GFP (green) and stained for Ac (red) and DEcad (blue). I, intermediate row; L, lateral row; M, medial row. Medial proneural clusters express Ac at early stage 7 (A, arrowheads), preceding lateral expression seen at the onset of stage 8 (B); Sna is robustly expressed in ingressed NBs during stage 9 (C). (D and E) Sna (green) is expressed by VM cells but not in Ac-positive medial proneural clusters (arrowheads) at stage 7 (D) and early stage 8 (E). (F) Ingressing NBs (asterisks) express Sna and enrich Ac at mid–stage 8. (G) Two medial NBs flanking the midline expressing Ac and Sna at stage 9. Bars: (A–C) 50 µm; (D and E) 10 µm; (F and G) 5 µm. (H) Mean apical area loss during NB ingression in mock-injected controls (blue) and embryos depleted for Sna, Esg, and Wor (red) by RNAi. (I) Mean apical area loss during NB ingression in Spider::GFP controls (blue) and shg^R69 mutants expressing ubi-DEcad::GFP (red). Data presented are means ± SEM.

Figure 3.

Apical constriction of NBs is achieved by cycles of strong contractions followed by short and weak expansions. (A) Single profile of apical membrane area loss during ingression. (B and C) Rate of apical area change in an NB (B) and a NIC (C) normalized to the apical area size at each time point. Positive rates are expansions, negative rates are contractions, and time 0 depicts the onset of ingression. (D) Relative area change (area(t) − area(t − 1 min)/area(t)) during contractions and expansions in NBs and NICs. Onset to end minus 10 min was considered early ingression, and the last 10 min was considered late ingression. n = 123–328 contractions/expansions per condition. From left to right: ****, P = 2.4 × 10⁻⁹; 5.3 × 10⁻⁷; 9.6 × 10⁻¹²; 4.7 × 10⁻¹⁷; and 2.4 × 10⁻²⁷; **, P = 0.005; ns (not significant), P = 0.97 and 0.33 (KS test). n = 19 NBs and 19 NICs (seven embryos). (E) Durations of contractions and expansions in seconds. Median values: 51.5, 39.4, and 28.8 s for contractions and 23.1, 14.2, and 32.2 s for expansions. From left to right: ****, P = 3.6 × 10⁻⁶; 6.6 × 10⁻¹²; and 1.7 × 10⁻¹⁰; **, P = 0.003; ns, P = 0.11 and 0.75 (KS test). Bars are IQRs. n values are as in D. (F) Fraction of time NBs and NICs spend contracting and expanding their apical membranes. From left to right: ****, P = 1.2 × 10⁻⁹; 1.4 × 10⁻¹¹; and 1.1 × 10⁻¹¹; ns, P = 0.155 (KS test). n values are as in D. (G) Mean period of apical oscillations in NBs and NICs; 183.6 ± 48.9 s for NBs and 161.7 ± 58.9 s for NICs. P = 0.25 (KS test). Data presented are means ± SEM. n values are as in D.

Figure 4.

Planar-polarized myosin distribution correlates with planar disassembly of NB cell contacts. (A) Stills of a time-lapse video of an NB (asterisk) labeled with Sqh::GFP (myosin, green) and GAP43::mCherry (cell membrane, red). Myosin is enriched at AP junctions (arrowheads). Time point 20 min corresponds to a 2.5-µm² apical area in NBs. Bar, 5 µm. P, posterior, A, anterior, D, dorsal, V, ventral. (B) Ratios between myosin intensity at AP (75–90°) and DV (0–15°) edges in NBs (n = 13 cells, seven embryos; data presented are means ± SEM) and in ectoderm cells during ingression. Time points are as in A. (C) Kymographs of myosin (Sqh::GFP) during the last 15 min of AP and DV edge disassembly in NBs. DV edges show more discontinuous myosin distribution than AP edges. Time intervals, 15 s. (D) Myosin levels (IQRs) in AP junctions (n = 12) and DV junctions (n = 11) in NBs versus DV junctions in NICs (n = 10). From left to right: ****, P = 2.5 × 10⁻¹⁷³; 2.8 × 10⁻²¹⁹; and P = 2.7 × 10⁻¹³ (KS test). (E) Correlation coefficients between AP/DV edge length and junctional myosin levels during ingression. Median correlation coefficient for AP edges was R = −0.54 and for DV edges was R = −0.39. Bars are IQRs. n = 20 junctions per condition, nine embryos. (F) AP and DV NB edge length during ingression. t = 0 min, onset of ingression. Speed of AP and DV edge contraction in NBs (means ± SD): −0.72 ± 0.3 µm/min and −0.33 ± 0.1 µm/min, respectively. P = 5 × 10⁻⁵ (two-tailed t test). Speed of DV edge length in NICs: 0.05 ± 0.06 µm/min. n = 11 junctions per condition, two embryos. (G–I) Representative plots of junctional myosin levels and AP or DV edge length in an NB and NIC. a.u., arbitrary units.

Figure 5.

Oscillations in medial and junctional myosin levels drive progressively stronger contractions in NBs. (A) Cycles of medial myos­in assembly and disassembly during early and late NB ingression. The cell membrane is labeled in red (GAP43::mCherry) and myosin in green (Sqh::GFP). Dashed circles track a pulse of medial myosin II assembly and disassembly during early and late ingression. (B and C) Representative plots of medial myosin levels versus apical area (B) and rates of medial myosin and area change (C) during ingression. (D and E) Myosin data are shifted backward and forward in time to calculate cross-correlations between changes in medial or junctional myosin and cell area during ingression. R₀ is the correlation coefficient with no time shift. A typical result is shown in E. (F and G) Maximum cross-correlation coefficients (F) and time shift distributions (G) as indicated in D and E. n = 63 NBs, 15 embryos. IQRs for medial myosin were −0.51, −0.45, and −0.34 and for junctional myosin were −0.47, −0.38, and −0.27. *, P = 0.01 (two-tailed t test). (H and I) Mean medial and junctional levels of myosin in an NB (H) and a temporally matched NIC (I). (J) Mean medial and junctional myosin levels during early ingression (up to last 10 min), late ingression (last 10 min), and in temporally matched NICs. n = 22 cells per condition. ***, P = 0.001; ****, P = 2 × 10−⁹; ns (not significant), P = 0.5 (left) and 0.7 (right; two-tailed t test). Data presented are means ± SD. a.u., arbitrary units. (K and L) Expansion of the apical cortex after laser ablation in NBs and NICs in endo-DEcad::GFP embryos. Maximum relative apical expansion (vertical dashed line in L): 97.0 ± 8.6% for NBs (n = 39) and 41.6 ± 3.3% for NICs (n = 36). P = 2.28 × 10⁻⁷ (two-tailed t test). Data presented are means ± SEM. (A and K) Bars, 5 µm.

Figure 6.

Medial myosin flows and fuses with AP and DV edges during NB ingression. (A) Classes of medial myosin behavior. Myosin clusters can assemble and coalesce at the cell center (>1 µm away from cell edges) or near (<1 µm) an AP or DV edge and either fuse (second and fourth rows) or not fuse (third and fifth rows) with that edge. Arrowheads track the movement of medial myosin clusters during NB ingression. Bars, 3 µm. (B) Planar distribution of medial myosin clusters. NBs show uniform planar distribution of myosin at the apical cortex, whereas medial myosin is more biased to flow parallel to the AP axis in NICs. P, posterior (0°), A, anterior (180°), D, dorsal (90°), V, ventral (270°). (C) Quantification of medial myosin behavior in NBs and NICs as described in A. n = 268 and 512 myosin clusters in NBs and NICs, respectively (19 cells per condition; seven embryos).

Figure 7.

Myosin controls the ratcheting behavior during apical constriction of NBs. (A) Apical area loss during ingression in videos of Spider::GFP-labeled control embryos (overexpressing myristoylated GFP; 28 NBs, four embryos) and zip-RNAi embryos (26 NBs, three embryos). (B) Apical area loss during ingression in ubi-DEcad::GFP-labeled wild-type (control, 26 NBs, four embryos) and sqh^M/Z mutants (17 NBs, three embryos). Two classes of mutant cells were observed, one with normal ingression speed (9/17 cells) and another significantly delayed group (8/17 cells). Data presented are means ± SEM. (C and D) Rate of apical area change during ingression of a control (C) and a delayed sqh^M/Z mutant NB (D). (E) Stills from time-lapse videos of NBs (segmented in green) in zip-RNAi and sqh^M/Z delayed NBs labeled with Spider::GFP and ubi-DEcad::GFP, respectively. See Fig. 1 F for control NB. Bars, 5 ­µm. (F) Fractions of time NBs spend contracting and expanding apically during ingression in control and sqh^M/Z class 2 NBs. **, P = 0.0078 (two-tailed t test). Data presented are means ± SD. (G) Period of apical oscillations in control (median, 120 s; n = 152 periods, 26 NBs) and sqh^M/Z delayed cells (median, 90 s; n = 433 periods, eight NBs). Box plot with IQRs and minimums/maximums displayed. ****, P = 3.2 × 10⁻⁹ (KS test). (H and I) Amplitude and duration of contractions and expansions in control and sqh^M/Z delayed NBs. Median amplitudes (bars; squared micrometers/minute): contractions, 3.88/3.14; **, P = 0.005; expansions, 1.75/1.99; ns (not significant), P = 0.11 (H). Median durations (s): contractions, 90/45; ****, P = 5 × 10⁻¹⁹; expansions, 45/30; ****, P = 1 × 10⁻⁷ (KS test in H and I). 126–393 events per condition.

Figure 8.

Canoe is required for NB ingression. (A) Stills from time-lapse videos of ingressing NBs in water-injected Sqh::GFP GAP43:mCherry embryos (control) and canoe-RNAi embryos. Arrowheads indicate the detachment of junctional myosin from an NB edge during ingression. Bars, 5 µm. (B) Apical area loss during ingression in control (22 NBs, four embryos) and canoe-RNAi embryos (24 NBs, three embryos). Data presented are means ± SEM. (C) Fractions of time ingressing NBs spend contracting and expanding in control and canoe-RNAi. *, P = 0.037 (KS test). Data presented are means ± SD. (D) Amplitude and duration of individual contractions/expansions during ingression in control and canoe-RNAi. Median amplitudes for control/RNAi (bars; squared micrometers/minute): contractions, 3.38/2.12; ***, P = 3 × 10⁻⁴; expansions, 1.15/1.08; ns (not significant), P = 0.599. Median duration for control/RNAi (seconds): contractions, 61.57/50.71, *, P = 0.012; expansions, 17.17/19.04, ns, P = 0.385 (KS test); n = 154–297 events per condition. (E) Amplitude and duration of apical contractions and expansions in NICs in control and canoe-RNAi embryos. Median amplitudes for control/RNAi (bars; squared micrometers/minute): contractions, 3.09/2.38; **, P = 0.005; expansions, 3.94/2.44; ****, P = 7.2 × 10⁻⁷. Median duration for control/RNAi (seconds): contractions, 33.35/33.73; ns, P = 0.983; expansions, 38.12/35.45; ns, P = 0.3613 (KS test). n = 275–401 events per condition.

Figure 9.

Cell contact and apical expansion of NICs adjacent to NBs are required for normal ingression. (A) NICs (purple) next to NBs expand apically 1.56 ± 0.47–fold during the last 25 min of ingression, in contrast to temporally matched NICs not in contact with NBs (1.05 ± 0.34). n = 25 cells per condition. **, P = 9.8 × 10⁻³ (KS test). (B) Mechanical uncoupling of NBs from neighbors by laser ablation increases ingression speed (top panels and plot) compared with sham-irradiated controls (bottom panels and plot). Each line represents the apical area of one NB after laser/sham cut. Time 0 depicts the apical area before cut. n = 8–9 cells per condition. (C) Cluster of NBs in Delta-RNAi embryos expressing Sqh::GFP (green) and GAP43::mCherry (red). Bottom panels show the same cluster with pseudocolored cells (yellow, blue, and red) ingressing sequentially. Myosin coalesces into foci (arrowheads) during late ingression. (D) Mean apical area loss in control (water-injected Sqh::GFP GAP43::mCherry; 22 NBs, four embryos) and Delta-RNAi embryos (58 NBs, three embryos). (E and F) Amplitude and duration of apical contractions and expansions during ingression in control and Delta-RNAi embryos. Median amplitudes for control/RNAi (E, squared micrometers/minute): contractions, 3.37/2.78; ns (not significant), P = 0.08; expansions, 1.15/1.31; ns, P = 0.509. Median durations for control/RNAi (F, seconds): contractions, 61.57/41.24; **, P = 0.001; expansions, 17.7/17.8; ns, P = 0.6 (KS test). 154–874 events per condition. (G) Cluster of ingressing NBs in Notch-RNAi. Cells are outlined by ubi-DEcad::GFP. DEcad is down-regulated between neighboring NBs (segmented in bottom panels). Bars, 5 µm. (H) DEcad::GFP levels at NB–NB junctions (Notch-RNAi) decrease during ingression, in contrast to NB–NIC junctions in controls. n = 15–16 junctions per condition. Data presented are means ± SEM. a.u., arbitrary units.

Figure 10.

Model of NB ingression. A typical example of the organization of junctional and medial myosin (red) during five stages of progressive apical constriction in an NB is illustrated. During early ingression (stages 1–3), medial myosin in the NB and neighboring cells flows toward AP NB edges, which become myosin enriched and disassemble. During late ingression (stages 4–5), progressively stronger pulses of medial myosin fuse with DV NB edges, which subsequently disassemble.