The formation of the MSxx-specific F-actin-rich structures is Arp2/3 dependent. (A-E) Images from movies of wild-type and Arp2/3 RNAi embryos expressing GFP::MOE. Wild-type embryos formed F-actin-rich structures (yellow arrowheads), whereas Arp2/3 RNAi embryos cell membranes did not (black arrowheads) and instead formed membrane blebs (asterisks, n=6/6). (F) Quantification of GFP::MOE ratios at the front and back MSxx cell boundaries in four Arp2/3 RNAi embryos, with color-coded GFP::MOE ratios as in Fig. 6. The preponderance of gray indicates that ratios are frequently similar in Arp2/3 RNAi embryos.

Arp2/3-depleted embryos localize apical PAR-3 and basolateral PAR-2 proteins normally. (A,B) Fixed embryos at gastrulation stage immunostained for PAR-2::GFP. In wild-type (n=12/12) and in Arp2/3 RNAi embryos (n=12/12), PAR-2::GFP localized basolaterally (arrowheads). (C,D) Wild-type embryos (n=7) and Arp2/3 RNAi embryos (n=6/6) immunostained for endogenous PAR-3 showed apical accumulation in all cells, including Ea/p (arrowheads). (E) anti-PAR-3 (red, n=7) and PAR-2::GFP (green, n=6) fluorescence intensity levels quantified in wild-type and Arp2/3 RNAi Ea/p cells show peaks of fluorescence intensity at the apical and basolateral membrane, respectively. Shading indicates 95% confidence intervals.

Dynamic, MSxx-specific, F-actin-rich extensions require Arp2/3 and apical PAR protein. Six cells converge in a ring (light gray), filling in the space left by the internalizing E cells (white gap in middle). Three of these cells are MS descendants. F-actin-rich structures formed in MSxx cells that contacted Ea/p cells, at their apical sites of contact (blue). These structures were absent in embryos deficient in either Arp2/3 or the apical marker par-3.

The Arp2/3 complex is enriched near cell membranes. (A-F) Embryos immunostained with ARX-5 antibodies (green). (A,B) Embryos at the 6- to 8-cell stage. ARX-5 appears enriched near plasma membranes (arrowheads). In control arx-5 RNAi embryos, enrichment near plasma membranes is reduced or absent. (C-F) Lateral views (C,D) and ventral views (E,F) of gastrulation-stage embryos also revealed ARX-5 localization near membranes (arrowheads). ARX-5 was similarly enriched at borders between MSxx and Ea cells (arrow). Cortical staining was absent in arx-5 RNAi embryos. Antibodies to P-granules were used to confirm permeabilization of embryos to immunostaining reagents (red). Nuclei are stained with DAPI (blue).

The formation of MS-specific F-actin-rich structures requires the apical PAR protein PAR-3. (A-E) Images from a movie of par-3 ZF1; GFP::MOE. An MSxx cell is labeled, and the Ea/p cells are labeled with asterisks. Ea/p cells remained in the plane of imaging as expected from an internalization defect reported in par-3 mutant embryos (Nance and Priess, 2002). Arrowheads indicate a boundary between MSxx and Ea cells. F-actin-rich structures did not form here in par-3 ZF1 (n=7/7). (F) Kymograph of a line across MSxx. The kymograph shows that the MSxx front cell boundary did not transiently accumulate GFP::MOE above levels seen at the back end. (G) Cortical ARX-5 localization (arrowheads) in a par-3 ZF1 mutant embryo.

Apical accumulation and activation of myosin is normal in Arp2/3-depleted embryos. (A-D) Wild-type and Arp2/3-depleted NMY-2::GFP embryos of the same age placed side-by-side show apical myosin accumulation (arrowheads). Arp2/3 RNAi embryos accumulated NMY-2::GFP apically in Ea/p (n=21/21 embryos). (E,F) Kymographs of the same embryos over 30 minutes. Arrowheads at the sides of the kymograph indicate the initial and final position of the NMY-2::GFP-enriched apical cortex. d, distance in μm from an arbitrary point. (G) Graph of cortical to cytoplasmic ratios of NMY-2::GFP fluorescence intensities over time in wild-type and Arp2/3 RNAi embryos. All measurements from wild-type (n=3) and Arp2/3 RNAi embryos (n=4) were plotted and lines that represent the averages of 5 minute intervals were drawn. Shaded regions indicate 95% confidence intervals. (H,I) Wild-type embryos stained with rMLC-P (p-rMLC) antibody shows apical distribution in Ea/p-enriched cells (white arrowheads) compared with neighboring cells (black arrowheads) in both wild-type and Arp2/3 RNAi embryos. Nuclear staining is a background signal (Lee et al., 2006).

Arp2/3 is required for stable membrane-cytoskeletal linkages at free cell surfaces. (A-H) Ventral views of embryos expressing PH::mCherry plasma membrane marker. (A) A surface view of a wild-type embryo. (B) A higher magnification of the Ep cell. The arrow indicates an apparent membrane tether. (C) A surface view of an Arp2/3 RNAi embryo. (D) A higher magnification of the Ea cell with blebs (arrowheads) and apparent infolding of the surface in the center of the cell (arrow). (E) A mid-plane view of a wild-type embryo. (F) A higher magnification of the Ea cell. (G) A mid-plane view of an Arp2/3 RNAi embryo. Blebs (arrowheads) formed at free apical surfaces. (H) A higher magnification of the Ep cell. Membranes at cell-cell contacts appeared normal (arrows). Blebs (arrowheads) formed only at free apical surfaces (n=37 embryos). (I) Images of an embryo expressing GFP::MOE and PH:mCherry. For PH::mCherry, asterisks indicate the membrane bleb. For GFP::MOE, asterisks indicate when F-actin accumulated beneath the bleb. (J) A 45-second kymograph of the images in I, showing individual markers and both markers merged. GFP::MOE is not enriched under the plasma membrane during bleb formation (red arrowhead). GFP::MOE then appears enriched near the plasma membrane (green arrowhead) and the membrane retracts (yellow arrowhead).

Arp2/3 depletion results in only partial shrinking of the Ea/p cell apical surfaces. (A-D) The Ea/p cells move into the interior of the embryo as surrounding cells fill in the gap left in the ventral (bottom) side. The Ea/p cells subsequently divide (D) in the interior. Only three of the four Ea/p descendants (shaded in purple) are marked with asterisks because the fourth descendant is not in same imaging plane. (E-H) Ea/p cells in Arp2/3-depleted embryos begin to move to the interior but fail to complete internalization and divide on the surface of the embryo. The progeny of Ea/p did eventually internalize as four cells. In this and subsequent figures, asterisks indicate Ea/p cells and/or Ea/p descendants, except where noted. (I) Lengths of exposed Ea/p apical surfaces are shown as ratios of the initial lengths ±95% confidence intervals. In wild-type embryos (n=6), Ea/p cells internalized by ∼14 minutes after MSa/p division. (J) In Arp2/3 RNAi embryos (n=6), the lengths of exposed apical Ea/p cell surfaces failed to completely decrease to zero. (K) Tracings of en face (ventral) views of individual cells. The apical surfaces of Ea/p cells in wild-type and Arp2/3 RNAi embryos were traced at 50 second intervals from films of the PH:mCherry membrane marker, and individual tracings were color-coded by time and overlaid. The colored scale indicates seconds after MSa/p division.

Three of the six cells surrounding the Ea/p cells produce dynamic, F-actin-rich structures. (A-C) Embryos viewed from their ventral sides. (A,B) GFP::MOE-expressing embryos. F-actin-rich structures (yellow arrowheads) formed at the borders of the MS descendants and the Ea/p cells. The AB descendants, ABplpa and ABplpp, did not form these F-actin-rich structures at their borders with the Ea/p cells (black arrowheads). F-actin enrichment at the P₄/Ep cell boundary is marked by an `x'. (C) An embryo expressing PH::mCherry showing F-actin-rich processes specifically at the borders of MS descendants and the Ea/p cells and not at the borders of AB descendants and the Ea/p cells. (D) Heatmap representing GFP::MOE accumulation in seven individual embryos. Yellow and blue colors indicate GFP::MOE front to back end ratios above 1.3-fold difference. Front and back ends are defined with respect to direction of extension across the gap as indicated in the diagram. The preponderance of yellow in the resulting heatmap indicates frequent enrichment of GFP::MOE at the front end. (E-H) Representative images from a wild-type GFP::MOE embryo, embryo 7 of the heatmap. (E) Mid-plane view. Ea/p cells are marked by asterisks and the three MSxx cells are labeled. (F-H) GFP::MOE accumulated at the border of MSxx and the Ea/p cells at certain times (yellow arrowhead) and not at other times (black arrowhead). We did not observe GFP::MOE accumulation at the back end of MSxx cells in the imaging plane shown, nor in other planes. (I) Kymograph of a line across an MSxx cell of a GFP::MOE expressing embryo. The front end of MSxx has dynamic enrichment of GFP::MOE. (J) Kymograph of a line across an ABxx cell of a GFP::MOE expressing embryo. The front end of ABxx does not have dynamic GFP::MOE enrichment.