RhoGAP RGA-8 supports morphogenesis in C. elegans by polarizing epithelia through CDC-42

CDC-42 regulation of non-muscle myosin/NMY-2 is required for polarity maintenance in the one-cell embryo of C. elegans. CDC-42 and NMY-2 regulate polarity throughout embryogenesis, but their contribution to later events of morphogenesis are less understood. We have shown that epidermal enclosure requires the GTPase CED-10/Rac1 and WAVE/Scar complex, its effector, to promote protrusions that drive enclosure through the branch actin regulator Arp2/3. Our analysis here of RGA-8, a homolog of SH3BP1/Rich1/ARHGAP17/Nadrin, with BAR and RhoGAP motifs, suggests it regulates CDC-42, so that NMY-2 promotes two events of epidermal morphogenesis: ventral enclosure and elongation. Genetic and molecular data suggest RGA-8 regulates CDC-42, and the CDC-42 effectors WSP-1 and MRCK-1, in parallel to F-BAR proteins TOCA-1 and TOCA-2. The RGA-8-CDC-42-WSP-1 pathway enriches myosin in migrating epidermal cells during ventral enclosure. We propose TOCA proteins and RGA-8 use BAR domains to localize and regenerate CDC-42 activity, thus regulating F-actin levels, through the branched actin regulator WSP-1, and myosin polarity through the myosin kinase MRCK-1. Regulated CDC-42 thus polarizes epithelia, to control cell migrations and cell shape changes of embryonic morphogenesis. Summary RGA-8, a protein with membrane binding and actin regulatory motifs, promotes embryonic morphogenesis by localizing active CDC-42 in developing epithelia, thus controlling actin and actin motors during cell movements.


Introduction
Organ and tissue formation are highly regulated processes during embryonic development. During embryogenesis, epithelial cells must develop and maintain apicobasal polarity and healthy cell-cell junctions as they move past or over other tissues in the process of morphogenesis. Defects in this process can lead to birth defects, or premature death. Epidermal morphogenesis in C. elegans can be divided into several stages. Epidermal cells are born at the posterior and dorsal surface of the embryo and become arranged in three types of adjacent cells with distinct behaviors -two rows of dorsal cells, and on each side of the embryo, two rows of lateral seam cells, and two rows of ventral cells, (Fig. 1A).
Epidermal morphogenesis begins when two outer rows, right and left ventral epidermal cells, migrate ventrally, pulling along other epidermal rows to enclose the embryo in a process known as ventral enclosure. This tissue migration is led by the two anterior-most ventral cells on each side, the leading cells. The leading cells reach the ventral midline first, while the more posterior ventral cells, the pocket cells, undergo a purse-like constriction, to enclose the embryo on the ventral side. Simultaneously, the two dorsal rows undergo dorsal intercalation to generate a single row of cells, in a process analogous to vertebrate convergent extension.
Epidermal elongation, the process that squeezes the embryo into a worm shape, begins when the two rows of lateral seam cells increase their length along the anterior to posterior axis, and decrease along the dorsal/ventral axis. The dynamics of actin and actomyosin contractility that regulate epidermal morphogenesis are regulated by the Rho GTPases.
In C. elegans, the three main GTPases, Rac1/CED-10, CDC-42 and RhoA/RHO-1, have been shown to be involved in some aspects of epidermal morphogenesis. Ventral enclosure is regulated by the GTPase Rac1/CED-10, which activates WAVE/Scar, a nucleation promoting factor (NPF) that activates branched actin formation through Arp2/3 [4,5,6]. Rac1/CED-10 and WAVE/Scar also regulate dorsal intercalation ( Removing LET-502/ROCK caused defects in elongation of the embryo. Activation of myosin II is achieved mainly through the LET-502/Rho kinase, but two additional kinases can contribute to maintain myosin II activity. P21-activated kinase PAK-1 and CDC-42-activated kinase MRCK-1 act in parallel to LET-502, since their loss enhances let-502 mutant severity.
MRCK-1 acts upstream of MEL-11 ( [14]  and is under the regulation of CDC-42 [15]. The role actomyosin contractility plays during ventral migration is not clear. ANI-1, a multidomain protein that organizes actomyosin contractility, is required to ensure proper alignment of contralateral leading-edge epidermal cells meeting at the ventral midline.
However, expression of ANI-1 was not detected in the epidermal cells, leading to the suggestion that ANI-1's action may be due to the interaction between the underlying neuroblast (neuron precursors) with the migrating epidermal cells. Interestingly, loss of rho-1, let-502 or mel-11 led to ventral enclosure defects, but it was not clear if this was due to roles in the neuroblasts or in the epidermis (

RGA-8 is a GAP for CDC-42 that regulates C. elegans embryonic morphogenesis
To characterize the regulators of CDC-42 during epidermal morphogenesis, we screened the C. elegans family of GTPase activating proteins (GAPs), which inactivate and double stranded RNAi to L1 worms and waiting 2 days. Under these conditions the RNAi knockdown resulted in approximately 70% embryonic lethality in their progeny, yielding the highest embryonic death during ventral enclosure stage at 61%, followed by elongation stage at 10%, and early differentiation at only 2.6% ( Fig. 1 A-D, Table 2). We hypothesized that removing GAPs from cdc-42 RNAi, which creates a hypomorphic cdc-42 background, might rescue embryonic lethality. We focused on one candidate CDC-42 GAP, named Rho-GTPase activating protein-8 (RGA-8). We generated an rga-8 deletion allele, pj60, using CRISPR ( Fig. 1D) and found it rescued cdc-42 RNAi embryonic lethality from 67% to 46% (Fig. 1E).
rga-8 was thus a candidate GAP for CDC-42.
To investigate which step of embryogenesis is regulated by RGA-8 and the CDC-42 pathway, we crossed mutants or fed RNAi food to dlg-1::gfp which is expressed at C. elegans apical junctions, and thus allowed us to monitor cell migration and cell shape changes. dlg-1::gfp; rga-8(pj60), resulted in 4% embryonic lethality, with approximately half the embryos arresting during ventral enclosure, and the other half during elongation (Tables 1,2). Another rga-8 allele, ok3242, a 611 bp deletion that causes a frame shift and premature stop codon, truncating RGA-8 part way through the GAP domain (Fig. 1D), showed higher embryonic lethality, at 7% (8.4% in dlg-1::gfp background) (Fig. 1C,E). We inactivated the GAP domain of RGA-8 by creating a CRISPR allele, pj71, which switches the catalytic Arginine to an Alanine, and measured 13% embryonic lethality, with arrests during ventral enclosure and elongation ( Fig. 1C,D,E). Thus, rga-8 has a role in two steps of epidermal morphogenesis, and its RhoGAP domain appears to be involved in this process.
Collectively, we suggest RGA-8 is a GAP for CDC-42, since putative null deletion mutations and RNAi depletion show genetic interactions with components of CDC-42 pathway signaling, including suppression of cdc-42 lethality.

RGA-8 is expressed in most cells with preferred localization at apical pharyngeal and intestinal cells.
To examine the expression pattern of RGA-8, we endogenously-tagged it at the Nterminus with mKate2, a red fluorophore, using CRISPR technology (

RGA-8 localization relative to CDC-42 biosensor in epidermis:
To test if RGA-8 regulates active CDC-42 in the epidermis, we first examined the gbd-wsp-1::gfp pattern at the ventral epidermis during enclosure. gbd-wsp-1::gfp is enriched at cell boundaries, but we did not detect obvious enrichment at the ventral midline. We examined the lateral seam cells that drive the first part of epidermal elongation. Pcdc-42::gbd-wsp- showed elevated epidermal F-actin levels (Fig. 5A, B). toca-2; toca-1 double mutants did not significantly alter epidermal F-actin levels. Therefore, a proposed CDC-42 pathway that includes RGA-8 and WSP-1 appears to be required to maintain appropriate F-actin levels in the migrating epidermal cells.

Protrusions dynamics are not significantly affected by the CDC-42 pathway
Ouellette and colleagues showed decreased dynamics in wsp-1 mutants, measured as how much the leading-edge membrane is displaced in the leading cells ( [21]Ouellette et al., 2016), although they did not examine F-actin levels. In contrast, we examined the dynamic formation of protrusion and retractions at the leading edge. We detected no significant change, though the number of protrusions was slightly increased in wsp-(gm324) and rga-8(pj60) mutants ( Fig. 5C,D). Therefore, the elevated F-actin levels seem to not significantly perturb dynamics.

Speed of ventral enclosure is affected by the CDC-42 pathway.
To test how changes in F-actin were affecting morphogenesis, we used the speed of migration as a phenotypic readout. To monitor timing for ventral enclosure using the epidermal F-actin movies, we measured from the time of the first protrusion to the first meeting at the ventral midline for the contralateral leading cells. rga-8(pj60) epidermal leading-edge cells migrate at slower rate (48min) compared to the wild type (24min), while wsp-1(gm324) and the toca-2;toca-1 showed slight delays relative to controls that were not statistically significant.

RGA-8 and CDC-42 pathway regulate NMY-2/myosin in epidermal cells.
While CDC-42 has been connected to ventral enclosure, it is not clear which effectors of cdc-42 loss via RNAi caused variable changes in the level of NMY-2/myosin II in the migrating pocket epidermal cells, perhaps due to the complex effects of depleting cdc-42. In contrast, wsp-1(gm324) pocket cells consistently showed a 20% reduction compared to wild type. Similarly, the rga-8(pj60) putative null mutant caused a 22% reduction compared to wild type. Interestingly, removing both TOCA-1 and TOCA-2 did not significantly change NMY-2/myosin II enrichment, while a triple mutant that removes all three, toca-2(ng11); toca-1(tm2016) rga-8(pj60), restored the level of NMY-2/myosin II back to wild type levels. A small deletion of rga-8, ok3242, predicted to truncate the C terminus, caused a 30% increase in myosin II/NMY-2, while the GAP point mutation, rga-8(pj71) caused a 15% drop, more similar to complete loss of rga-8 (Fig. 6C). Measurements of myosin levels in the migrating leading cells showed similar changes, with approximately 20% drop in rga-8(pj60) (not shown). Thus, RGA-8 and CDC-42 pathway proteins regulate accumulation of myosin II/NMY-2 in enclosing epidermal cells (Fig. 6).

RGA-8 and CDC-42 regulate the myosin light chain kinase, MRCK-1, in epithelia
To address how changes in cdc-42 pathway genes may result in altered myosin levels, we  it localized to specific subcellular membranes. We found it has diffused enrichment around apical epithelia, and may only be transiently enriched at epidermal membranes (Fig. 3).
Shared phenotypes with CDC-42 pathway: The phenotypes of three rga-8 alleles and of cdc-42 pathway components suggested a common function during epithelial morphogenesis. In particular, we found that deletion alleles of the toca-2; toca-1 double, or rga-8, or wsp-1, or cdc-42 RNAi, led to similar embryonic lethality with arrests during both ventral enclosure and elongation, and expanded intestinal apical lumen (Fig. 1B,C). Mutations in RGA-8 led to highly penetrant changes in epidermal cell migration timing, in levels of epidermal F-actin and in levels of epidermal myosin (Fig. 5, 6). Similar changes were also seen in mutations of toca-2;toca-1, wsp-1, and cdc-42 RNAi. Embryonic lethality was not fully penetrant for any of the RGA-8 alleles. This is also true for complete loss of toca-2; toca-1 (13%) or wsp-1 (35%).
One explanation for this ability of embryos to sometimes survive without the contribution of the TOCAs/RGA-8/ CDC-42/WSP-1 pathway is that this pathway is responsible for only part of the events of epidermal morphogenesis. In contrast, the CED-10/WAVE pathway leads to 100% embryonic lethality, suggesting that Rac/WAVE regulation of Arp2  (Fig. 7), and decreased nmy-2::gfp (Fig. 6), suggesting misregulated CDC-42 results in misregulated MRCK-1 that alters levels of myosin, and also, perhaps its localization. Higher resolution live imaging of myosin during these dynamics events can address this model. Further experiments will need to investigate which of the many cellular processes regulated by CDC-42 are promoted by the RhoGAP-BAR protein RGA-8. All worms were grown at 23°C unless otherwise indicated. Early arrest: includes embryos with cytokinesis defects and other early arrest so that tissue differentiation is defective. Lethal arrest occurs before epidermal morphogenesis begins.
Ventral enclosure: describes embryos with fully differentiated tissues that fail epidermal enclosure so that internal organs end up on the outside (AKA Gex, gut on exterior).
Elongation: describes embryos that enclose but then fail to lengthen along the anterior/posterior axis. Some of these burst after initially enclosing.

Strains -
All strains used in this study are listed in Table 3. Strains were either received from the CGC (Caenorhabditis Genetics Center, USA), the NBRP (National Bio-Resource Project, Japan), or individual labs listed below, or were generated for this study. All strains were grown at 23°C unless otherwise stated. injected with dpy-10 sgRNA. Silent mutation that inserted a new AvaI restriction enzyme site was engineered on the rescue oligos (MSo1602) to ease the screening process of identifying the mutants. DNA sequencing was done to verify the mutations and inserts for all strains. All strains were back crossed to wild type at least three times to minimize possible off-target effects by the gRNAs and Cas9. Primers used for strain generation are listed in Table 4.

RNAi experiments
All RNAi bacterial strains used in this study were administered by the feeding protocol as in

Myosin measurements
To compare myosin levels as pocket cells meet, a rectangular box enclosing the pocket cells as they first touch (approximately 320min) was drawn (yellow box in Fig. 6C), Time points shown and measured are those when the epidermal cells first touch based on epidermal Factin signal. To measure myosin puncta on the same focal plane as the epidermis, we co-localized the highest Plin-26::LIFEACT::mCherry intensity with NMY-2::GFP and recorded the highest of three measurements per embryo (Fig. 6B). Maximum intensity values were recorded after subtracting the average background fluorescence. The graph in Fig. 6B records the relative level of NMY-2::GFP, after normalization of either WT, or WT on control RNAi, to 1.