A. The Rac-WAVE/SCAR-Arp2/3 pathway.
A signaling pathway that connects migration signals or cues, through the small GTPase CED-10/Rac1, through the WAVE/SCAR nucleation promoting complex, to the Arp2/3 complex is thought to promote branched polymerization of actin. The known C. elegans homologs are shown. Plants and humans contain a fifth component of the WAVE/SCAR complex, HSPC300/BRICK (Eden et al., 2002; Frank et al., 2003; Le et al., 2006; Cascon et al., 2007) but homology searches have not identified a homolog in C. elegans.
B. C. elegans WVE-1/WAVE has similar domains to human WAVE.
WVE-1 shows a high degree of homology to all WAVEs and a conserved domain structure. C. elegans WVE-1/WAVE is 31% identical to human WAVE2 over its entire length, and shows similar homology to WAVE1 and WAVE3. Like other WAVEs it contains the Wave Homology Domain (WHD), including the basic region, a Proline-rich region thought to mediate profilin binding, and the verprolin homology, cofilin homology and acidic (VCA) region through which WAVEs are thought to bind actin and the Arp2/3 complex. The two wve-1 mutations, ne350 and zu469, are predicted to lead to early truncations of the protein.
C. Loss of wve-1, abi-1 or Arp2/3 complex components leads to the Gex phenotype.
Embryos from mothers injected with RNA corresponding to Arp2/3 components ARP-2/K07C5.1, ArpC5/M01B12.3 and ArpC4/C35D10.6 show the Gex-like epidermal enclosure defect. Animals produced early when RNAi is not fully penetrant still have maternal product so they display zygotic phenotypes including a Ced (cell death defective) corpse engulfment defect. The Ced phenotype is also seen in wve-1 genetic mutants and abi-1 RNAi embryos. The Ced phenotype is well studied for ced-10 and has been seen in gex-2 and gex-3 mutant embryos (Reddien and Horvitz, 2000; Kinchen et al., 2005; Soto et al., 2002). White arrows: anterior of the pharynx; white arrow heads: anterior of the intestine; black arrows: unengulfed apoptotic cells. Embryos are shown at a late stage (at least 700 minutes after first cleavage), when wild-type larvae have few unengulfed corpses (Soto et al., 2002). Embryos in all figures are oriented with anterior at left and dorsal up unless otherwise indicated. Scale bars in all figures represent 10um.

Larvae carrying the DLG-1::GFP transgene (Totong et al., 2007) were cultured from the first larval stage (L1) on either control food (no RNAi) or food containing gex-3 or arp-2 dsRNA. During larval stages (L2 shown here) epidermal seam cells in wild-type, gex-3(RNAi) and arp-2(RNAi) animals are unfused (white arrows indicate cell boundaries) (Podbilewizc and White, 1994). After the L4 to adult molt, wild-type (Ambros, 1989), gex-3(RNAi) and arp-2(RNAi) animals undergo seam cell fusions. The efficiency of post-embryonic RNAi was monitored by checking adults for the Gex Egl phenotype and their progeny for the Gex Mel phenotype. The slight twist in the seam cells is due to the dominant rol-6(su1006) mutation used to mark the DLG-1::GFP transgene (Totong, et al., 2007).

A. Live imaging with a filamentous actin-binding transgene.
A plin-26::vab-10 ABD (actin binding domain)::gfp transgene was used to investigate protrusion extension at the leading edge. This line is not integrated, so care was taken to analyze only those embryos that expressed the transgene evenly in the epidermis. Protrusion extension was analyzed over 25 minutes as described for Fig. 4B. Wild-type leading cells are up to 3.4 times as protrusive as the pocket cells when analyzed with this transgene. In contrast, the leading cells of arp-2(RNAi) embryos are only slightly more protrusive than the pocket cells.
B. Actin binding protein enrichment at the leading edge of leading cells requires ARP-2.
A plin-26::vab-10 ABD::gfp transgene was used to compare filamentous actin enrichment at the leading edge of wild-type and arp-2(RNAi) epidermal cells over a 20-minute period (while ventral enclosure is proceeding). Average fluorescence intensity within an equally sized region at the leading edge of the leading cells was compared. This transgene also illustrates changes in cell shape that occur in arp-2(RNAi) embryos.

A. The apical intestine is altered in Gex embryos.
Top row: An erm-1::gfp transgene (Gobel et al, 2004) is enriched apically in wild-type and mutant intestines, although the apical domain is wider in the mutants (split arrows). Second row: Phalloidin staining shows that wild-type embryos are enriched for F-actin at the apical region of the pharynx (anterior arrows), and intestine (split arrows). Dorsal arrows indicate the body wall muscles. Third row: Enlarged image of intestine (yellow box). In Gex embryos phalloidin levels (green) drop in the apical intestine. DAPI (red) identifies the intestinal nuclei.
B. Apical F-actin shows a pronounced drop in the embryonic intestine of Gex embryos.
The phalloidin signal, in the intestine of wild-type and Gex embryos, was measured using ImageJ (n=at least 12). Using the DAPI signal to confirm intestinal cell location, an equally sized area of each embryo was tested for average fluorescence intensity.
C. Enclosed wve-1(RNAi) animals show increased lumen area.
Wild-type animals and wve-1(RNAi) escaper animals that enclose and reach the 3-fold stage were compared. Intestine length (between arrows) and lumen width (brackets) were marked by the MH27 antibody to the junctional protein AJM-1 (n= at least 10).
D. Summary of the gex intestinal epithelial defects.
Cartoons of cross-sections through the intestine summarize apical/basal changes in Gex embryos. The lumen is shown in green, adherens junctions are shown in red and basolateral regions are shown in black. Loss of Arp2/3-dependent actin nucleation leads to an expansion of the apical domain, a basal shift of the junctional domains, and cell shape changes in the basolateral regions.

A. Gex terminal embryonic phenotypes. Left panel: DIC images. In wild-type embryos the pharynx (arrows) and intestine (arrow heads) are maintained internally by epidermal enclosure. In Gex embryos these tissues lie outside at the end of embryogenesis. Center panel: mAb MH46 to LET-805/Myotactin (Francis and Waterston, 1991; Hresko et al., 1999) shows that the pharynx (left arrows) is short and disorganized in Gex embryos while body-wall muscles (right arrows) fail to elongate in Gex embryos. Right panel: mAb MH33 to IFB-2 (Francis and Waterston, 1985; Bossinger et al., 2004) shows the width of the intestinal lumen (split arrows), which expands in Gex embryos. Embryos are pictured at late embryogenesis.
B. Morphogenesis defects in epithelial tissues of Gex embryos using DLG-1::GFP.
Comparison of wild-type embryos carrying the DLG-1::GFP transgene (Firestein and Rongo, 2001; Totong et al., 2007) and the same strain fed RNAi food for wve-1, gex-2, gex-3 or arp-2 shows disruption of morphogenetic movements. Top row: Dorsal intercalation fails in Gex embryos. Second row: In wild type (lateral view), cells of the intercalated dorsal row expand laterally, then fuse (asterisk) to form syncytial hyp7 (Podbilewicz and White, 1994). In Gex embryos (dorsal view) dorsal cells fail to expand laterally, yet they fuse as in wild type (asterisks). Third row: In wild-type embryos the ventral-most rows (open arrow) migrate and meet up at the ventral surface (Chin-Sang and Chisholm, 2000; Simske and Hardin, 2001). The ventral-most cells (open arrows) in Gex embryos arrest migration at or close to where they are born. Dotted lines indicate the unenclosed regions of each embryo. Fourth row: A wild-type embryo (ventral view) shows the epidermal cells meeting at the ventral surface and enclosed internal organs are not visible. Gex embryos (3/4 side view) show arrested epidermal cells and internal organs exposed on the ventral surface. Split arrows show width of intestinal DLG-GFP domain.
C. A summary of the morphogenesis defects caused by loss of Rac, WAVE/SCAR or Arp2/3. Loss of ced-10/Rac1, wve-1, gex-2, gex-3, abi-1 or arp-2 leads to the distinctive Gex (gut on the exterior) phenotype due to failures in cell movement and cell shape changes. The epidermis is shown in blue, the pharynx in green and the intestine in red. Lateral view: Gex embryos fail to initiate epidermal ventral movements. By 400 minutes after first cleavage, wild-type embryos initiate circumferential constrictions to squeeze the embryo into a worm. Gex embryos undergo constriction that leads the epidermis to collapse inwardly. The internal organs (pharynx and intestine) end morphogenesis exposed on the ventral surface by 700 minutes. Dorsal view: in Gex embryos dorsal intercalation fails (330 minutes), but this does not prevent fusion of the dorsal epidermis (by 450 minutes).

A. wve-1, gex-2, gex-3 and arp-2 mutants display similar defects of morphogenesis.
Time points were extracted from 180 minute 4D movies (270 to 450 minutes after first cleavage) of wild-type and Gex embryos (Supplemental Movie 1). The nhr-25::yfp transgene is expressed in both the nucleus and cytoplasm of epidermal cells, but not in the underlying cells of the embryo (dotted line). Wild-type embryos undergo dramatic morphogenetic changes during the two hours of development shown (from 280 to 400 minutes). The first epidermal movement detected with nhr-25::yfp is a ventral extension of the posterior epidermal cells, probably due to cell division along the anterior/posterior axis. This movement occurs in Gex embryos (left panel, arrows). The next movement detected is the ventral cell migration of epidermal cells to form an even ventral row. In Gex embryos ventral movements fail, leading to an uneven lateral row of cells (asterisks). In wild-type embryos the epidermal cells go on to enclose the embryo. In Gex embryos ventral epidermal cells display arrested migration, despite protrusions that lasts at least 5 hours. In wild-type embryos the epidermis undergoes circumferential constriction at 700 minutes while in Gex embryos epidermal cells collapse inwardly at the equivalent time point (Fig. 2B, 2C; unpublished movies). At least two movies were analyzed for each genotype. For the analysis below we chose movies where the Gex phenotype is confirmed by following the embryos for at least 6 hours of morphogenesis and made 4D QuickTime projections which permit viewing of the embryos from multiple angles (http://fsbill.vcell.uchc.edu/gloworm/Patel_et_al_2008_Data/). Measurements were done on one to two movies for each genotype. The movies were staged by DIC analysis and by noting the first appearance of the nhr-25::yfp transgene, at approximately 280 minutes after first cleavage.
B. The leading cells are more protrusive than the pocket cells in wild-type embryos. The ventral epidermal cells in a wild-type animal carrying the nhr-25::yfp transgene show how protrusions in leading cells (#2,3) vs. pocket cells (#4–8) were measured. The perimeter of the protrusive edge of the leading cells was traced using the Freehand Tool and the width of the leading cells was measured using the Segmented Line tool (ImageJ). Protrusion extension at the leading edge is represented as the ratio of the perimeter of the leading edge cells relative to their width. The same analysis was performed on the pocket cells. The analysis was performed on 5 time points, 10 minutes apart, spanning a 40-minute time interval when ventral enclosure occurs in wild type.
C. The ratio of leading cell to pocket cell protrusions changes in Gex embryos. Protrusions in leading cells and pocket cells were measured as above at five time points when ventral enclosure occurs in wild type. To compare the protrusion extenstion of the leading cells to that of the pocket cells we divided the protrusive extension of the leading cells by the protrusive extension of the pocket cells and plotted this ratio. The ratio of leading cell to pocket cell protrusions is shown for wild-type and Gex embryos.
D. The number of protrusions per cell per minute drops in Gex embryos. The number of protrusions made by leading cells and pocket cells during 17 time points 2.5 minutes apart (from 300 to 340 minutes) was counted in wild-type and Gex embryos.
E. Table of protrusion number comparisons. The number of protrusions made by the ventral epidermal cells (cells 1–8, Fig. 4B) was counted over a period of 40 minutes as in 4D. The average protrusions/cell/minute, relative to wild type, is shown for all cells (Total, #1–8), or for subsets of the cells (LC: leading cells, #2,3; PC: pocket cells, #4–8).

A. F- actin and WAVE/SCAR proteins are enriched in the basolateral epidermis and apical intestine.
GFP::WVE-1 and GFP::GEX-3 rescuing transgenes are expressed in the cytoplasm and enriched at membrane regions. This localization pattern resembles the enrichment of F-actin visualized with phalloidin. In epidermal cells, where apical is outside and basolateral is inside, GFP::WVE-1, GFP::GEX-3 and F-actin (green) are enriched basolaterally at the bean stage (380 minutes, top panels) as shown by their localization basal to the adherens junction associated proteins DLG-1 and AJM-1. In intestinal and pharyngial cells, where apical is inside and basolateral is outside, GFP::WVE-1, GFP::GEX-3 and F-actin are enriched at apical regions by the 1.5-fold stage (440 minutes, bottom panels). Arrows point to the developing nerve ring.
B. & C. WVE-1 and GEX-3 require other WAVE/Scar components for abundance.
The rescuing GFP::WVE-1 and GFP::GEX-3 transgenic strains were grown on control bacteria (no RNAi, WT) or bacteria expressing double stranded wve-1, gex-2, gex-3 or abi-1 RNA.
B. Embryos fed the control or WAVE/SCAR bacteria are shown. In all cases, loss of other WAVE/SCAR components leads to lower expression of GFP::WVE-1 and GFP::GEX-3. Remaining spots are produced by autofluorescent gut granules. DAPI staining indicates the stage (at least 500 minutes) and orientation (widely spaced intestinal nuclei are ventral).
C. The graph indicates the percentage of embryos expressing GFP when the GFP::WVE-1 and GFP::GEX-3 transgenic strains are grown on wild type bacteria or bacteria expressing WAVE/SCAR components.
D. GEX-2, GEX-3 and ABI-1 are required to maintain WVE-1 protein levels.
Embryonic lysates were depleted of wve-1, gex-2, gex-3 or abi-1 by feeding control bacteria or bacteria expressing wve-1, gex-2, gex-3, or abi-1 dsRNA to the RNAi sensitive strain rrf-3(pk1426). Western blots were probed with antibodies to WVE-1, GEX-2 and GEX-3. The effectiveness of the RNAi food was measured as percent embryonic lethality. The bottom of each gel was cut and probed with antibodies to tubulin as a loading control. 20ug of embryonic lysates were loaded per lane.