Identification of the pLLP transcriptional profile by scRNA-seq. (A) Unsupervised clustering and UMAP reduction diagram of cells derived from Tg(-8.0claudinB: lynGFP)^zf106; TgBAC(cxcr4b:F-tractin-mCherry)^p3 embryos. (B,C) Feature plots of hmx2 and hmx3a identify clusters 0, 7 and 9 as pLLP. (D) Unsupervised reclustering of clusters 0, 7 and 9. (E,F) Feature plots of ki67 and pcna identify cluster 2 as proliferating cells. (G,H) Feature plots of lef1 and notum1a identify cluster 1 as leader cells. (I,J) Feature plots of etv4 and ackr3b identify cluster 0 as follower cells.

Expression profile of genes that regulate actin dynamics. (A) Dot plot expression profile of mcf2lb, twf2b, arhgef4 and fhdc1. (B-E) Feature plots showing expression profile of the above four genes in pLLP clusters. (F-I) In situ hybridization of the four genes. Note that expression profiles via in situ hybridization largely match those observed by scRNA-seq. Arrowheads indicate rosette centers. Dotted lines indicate the outline of the pLLP. Scale bars: 10 μm.

Loss of Mcf2lb leads to the deposition of supernumerary NMs. (A,B) Tg(-8.0claudinB: lynGFP)^zf106 marks NMs in WT and mcf2lb mutant embryos at 3 dpf. Note the excess number of deposited trunk NMs in the mcf2lb mutants. Asterisks mark deposited trunk NMs. Full-body images in A were stitched together from four fields of view of individual body segments (head, trunk, tail) using the auto-blend feature in Photoshop. Full-body images in B were stitched together from three fields of view of the head and trunk using the auto-blend feature in Photoshop, and the tail image was manually included. (C) Mean number of NMs in WT (n=8 embryos) and mcf2lb mutant (n=12 embryos) embryos at 3 dpf. (D-E′) DAPI staining of nuclei in deposited NMs in WT (D,D′) and mcf2lb mutants (E,E′). (F) Mean number of cells per NM in WT (n=14 NMs from four embryos) and mcf2lb mutants (n=26 NMs from five embryos). (G) Mean number of cells in the first NM in WT (n=19 NMs) and mcf2lb mutants (n=25 NMs). ***P<0.001 (unpaired two-tailed t-test). n.s., not significant. Error bars are s.d. Scale bars: 100 μm (A,B); 5 μm (D,E).

mcf2lb mutants show abnormal organization of the pLLP and NM deposition. (A-B″) Lateral views of stills from time-lapse movies obtained between 30 and 44 hpf of either WT or mcf2lb mutants positive for Tg(-8.0claudinB: lynGFP)^zf106 (Movies 1,2). (C) Mean pLLP length in WT (n=7 pLLPs) and mcf2lb mutants (n=8 pLLPs). (D) Mean number of cells within the pLLP in WT (n=19 pLLPs) and mcf2lb mutants (n=19 pLLPs). (E) Mean velocity of the pLLP in WT (n=7 pLLPs) and mcf2lb mutants (n=6 pLLPs). (F-G″) High-magnification lateral views of still images from movies of the migrating pLLP in WT and mcf2lb mutants (Movies 3,4). pLLPs were imaged for ∼1.5 h starting at ∼30 hpf. Note the abnormal organization of the pLLP in mcf2lb mutants and the beads on a string appearance along the midline instead of distinct clusters of membrane as in WT (Movies 3,4). Asterisks indicate foci. Arrowheads indicate points. (H) Mean number of foci per pLLP in WT (n=15 pLLPs) and mcf2lb mutants (n=18 pLLPs). (I) Mean number of points per pLLP in WT (n=15 pLLPs) and mcf2lb mutants (n=18 pLLPs). (J) Scatter plot of foci numbers per pLLP versus number of points per pLLP in WT (n=15 pLLPs) and mcf2lb mutants (n=18 pLLPs). *P<0.05, **P<0.01 (unpaired two-tailed t-test). n.s., not significant. Error bars are s.d. Scale bars: 20 μm (A,B); 10 μm (F,G).

mcf2lb mutants show impaired apical constriction of cells incorporated into rosettes. (A,B) Cellular reconstruction of WT and mcf2lb mutant pLLPs. (C,D) Examples of apically constricted cells from WT (C) and mc2lb mutant (D) pLLPs. The left side panels show the lateral (top-down) view as in panels A and B, and the right panels show the apical/basal views of the cell (cell is virtually turned by 90°). Blue and purple cells marked by the asterisks in panels A and B are also shown in C and D, respectively. (E) Examples of cells that are making contact with multiple points in mcf2lb mutants. Left panel is a lateral (top-down) view; right panel is an apical/basal view. (F) Categorical breakdown of cells in the trailing region of WT (n=258 cells from four pLLPs) and mcf2lb mutants (n=311 cells from four pLLPs). Note that in mcf2lb mutants, there is an increase in the percentage of cells that touch multiple points and an increase in the percentage of cells not incorporated into rosettes. Subsequently, there is a decrease in the percentage of cells incorporated into rosettes. (G) Mean apical constriction index of cells that are incorporated into rosettes in WT (n=169 cells from four pLLPs) and mcf2lb mutants (n=139 cells from four pLLPs). (H) Mean cell volume of cells incorporated into the trailing rosette in WT (n=74 cells from four pLLPs) and mcf2lb mutants (n=91 cells from four pLLPs). (I) Mean apical width of cells incorporated into rosettes in WT (n=169 cells from four pLLPs) and mcf2lb mutants (n=139 cells from four pLLPs). (J) Mean basal width of cells incorporated into rosettes in WT (n=169 cells from four pLLPs) and mcf2lb mutants (n=139 cells from four pLLPs). Note that the apical width is significantly increased in mcf2lb mutants whereas there is no significant difference in the basal width. ***P<0.001 (F: chi-square test; G,I,J: Mann–Whitney U-test; H: unpaired two-tailed t-test). n.s., not significant. Error bars are s.d. Scale bars: 10 μm (A,B and E pLLP); 5 μm (C,D and E individual cells).

mcf2lb mutants show greater variability in apical membrane dynamics. (A-B‴) Donor cells derived from either WT or mcf2lb mutant Tg(prim:lyn2-mCherry) (magenta) were transplanted into WT or mcf2lb mutant Tg(-8.0claudinB: lynGFP)^zf106-positive (green) embryos, respectively, and mounted for live imaging between 30 and 32 hpf (Movie 5,6). Tg(prim:lyn2-mCherry) cells are shown in grayscale for clarity. Asterisks indicate cells used for analysis. (C) Mean membrane variability of transplanted WT (n=22 cells from seven embryos) or mcf2lb mutant (n=13 cells from three embryos) cells over time from movies obtained from experiments in panels A and B. Membrane variability is defined as the standard deviation of the apical width of a cell over an hour period of imaging in apically constricted cells. Error bar is s.d. (D) Apical membrane width over time with s.e.m. (shaded area). Note higher variability in the mutant. **P<0.01 (Mann–Whitney U-test). Scale bars: 10 μm.

pLLP cell polarity is largely unaffected in mcf2lb mutants. (A-J′) Immunostaining for the tight junction marker ZO-1 in WT and mutant pLLPs at 45 hpf. Panels A-J show the lateral (top-down) view, A′-J′ show an apical-basal view (images were digitally rotated 90° around x-axis; dashed lines in A,B,F,G). All images are z-projections. Dotted lines indicate the pLLP in B and F and NMs in B-E and G-J. Note the ring structure and the apical localization of ZO-1 in the trailing-most rosette and deposited NMs in WT. In contrast, ZO-1 signal is disorganized in mc2lb mutants. (K) Quantification of the percentage of apical- and basal-localized ZO-1 signal in the pLLP in WT (n=8 pLLPs) and mcf2lb mutants (n=13 pLLPs). (L) Quantification of the percentage of apical- and basal-localized ZO-1 signal in NMs in WT (n=15 NMs from four embryos) and mc2lb mutants (n=26 NMs from five embryos). (M-N′) Par-3-tagRFP expression in WT and mcf2lb mutant pLLPs. Panels M and N show the lateral (top-down) view, panels M′ and N′ show an apical-basal view (images were digitally rotated 90° around x-axis; dashed lines M,N). All images are z-projections. Note Par-3 localization to the rosette centers and the midline in WT embryos; in contrast, Par-3 is localized to the midline but not organized into rosette centers in mcf2lb mutants. Par-3 is apically localized in both WT and mcf2lb mutant pLLPs. (O) Quantification of the percentage of apically- and midline-localized Par-3 signal in WT (n=7 pLLPs) and mcf2lb mutants (n=8 pLLPs). ***P<0.001 (unpaired two-tailed t-test). n.s., not significant. Error bars are s.d. Scale bars: 10 μm (A,F,M,N); 5 μm (B-D,G-J).

RhoA signal is lost at rosette centers in mcf2lb mutants. (A,B) Fluorescent signal from the RhoA sensor line TgBAC(cxcr4b:AHPH-GFP) at 36 hpf in WT (A) and the mcf2lb mutant (B). Tg(prim:lyn2-mCherry) marks cell membranes in the migrating pLLP. Dashed lines mark the pLLP. Arrowheads indicate rosette centers and points. Note the absence of RhoA signal at the rosette centers and points in the mcf2lb mutant pLLP. (C) Quantification of RhoA signal at the center of the trailing-most rosette (point) within the migrating pLLP in WT (n=13) and mcf2lb mutants (n=10). ***P<0.001 (unpaired two-tailed t-test). Error bars are s.d. Scale bars: 20 μm.

RhoA signaling is disrupted in mcf2lb mutants. (A-B′) Immunostaining for Rock2a in WT and mcf2lb mutants. A,B show the lateral view (top-down); A′,B′ show the apical-basal view (images were virtually rotated 90° around the x-axis; dashed lines A,B). All images are z-projections. Images are masked to show signal within the pLLP. Note the localization of Rock2a is diminished in mcf2lb mutant pLLP. (C) Total fluorescence of Rock2a per cell in WT (n=10 pLLPs) and mcf2lb mutants (n=10 pLLPs). (D) Fluorescence of Rock2a at rosette centers in WT (n=16 points from ten pLLP) and mcf2lb mutants (n=30 points from ten pLLPs). (E-F′) Immunostaining for pMRLC in WT and mcf2lb mutants. E,F show the lateral (top-down) view; E′,F′ show the apical-basal view (images were virtually rotated 90° around the x-axis; dashed lines E,F). All images are z-projections. Images are masked to show signal within the pLLP. Note diminished localization of pMRLC in the mcf2lb mutant pLLPs. (G) Total fluorescence of pMRLC per cell in WT (n=10 pLLPs) and mcf2lb mutants (n=13 pLLPs). (H) Fluorescence of pMRLC at the gatherings of the membranes in WT (n=17 points from ten pLLPs) and mcf2lb mutants (n=37 points from 13 pLLPs). Dotted lines indicate pLLP. **P<0.01, ***P<0.001 (unpaired two-tailed t-test). n.s., not significant. Error bars are s.d. Scale bars: 10 μm.