WASP-Arp2/3-dependent actin polymerization influences fusogen localization during cell-cell fusion in Caenorhabditis elegans embryos

ABSTRACT Cell-cell fusion is essential for development and physiology. Actin polymerization was implicated in the Caenorhabditis elegans fusogen EFF-1 engagement in a reconstituted Drosophila cell culture system, and the actin-binding protein spectraplakin links EFF-1 to the actin cytoskeleton and promotes cell-cell fusions in C. elegans larvae. However, it remains unclear whether and how fusogens and the actin cytoskeleton are coordinated in C. elegans embryos. Here, we used live imaging analysis of GFP knock-in and RNAi embryos to study the embryonic cell-cell fusions in C. elegans. Our results show that the inhibition of WASP-Arp2/3-dependent actin polymerization delays cell-cell fusions. EFF-1 is primarily distributed in intracellular vesicles in embryonic fusing cells, and we find that the perturbation of actin polymerization reduces the number of EFF-1-postive vesicles. Thus, the actin cytoskeleton differently promotes cell-cell fusion by regulating fusogen localization to the fusing plasma membrane in larvae or to intracellular vesicles in embryos.

The actin cytoskeleton regulates myoblast fusion in Drosophila, zebrafish and mice (Abmayr and Pavlath, 2012;Chen, 2011;Gruenbaum-Cohen et al., 2012;Kim et al., 2015;Sens et al., 2010;Shilagardi et al., 2013). In a reconstituted Drosophila S2R+ cell-cell fusion system, actin polymerization facilitates EFF-1 enrichment and engagement at invasive membrane protrusions (Shilagardi et al., 2013). Our recent results from the C. elegans postembryonic cell-cell fusions support the findings from the reconstituted system. In the epithelial seam and hyp7 cell-cell fusions of C. elegans larvae, EFF-1 and F-actin accumulate at the cortex of two fusing cells, and WASP-Arp2/3-dependent actin polymerization promotes this process by recruiting EFF-1 to the fusing plasma membranes (Yang et al., 2017). Importantly, we identified that the actin-binding protein spectraplakin/VAB-10A directly links EFF-1 to the actin cytoskeleton (Yang et al., 2017).
However, in C. elegans embryos, RNA interference of the actin nucleation factor Arp2/3 complex and the actin nucleation-promoting WAVE/Scar complex did not perturb embryonic epidermal cell fusion (Patel et al., 2008;Xiong et al., 2011). Using an EFF-1::GFP knock-in nematode, we showed that EFF-1 is primarily distributed to intracellular vesicles and may only transiently localize at fusion sites in embryonic fusing cells, which is consistent with the EFF-1 localization pattern uncovered using immunofluorescence and a functional GFP reporter (Smurova and Podbilewicz, 2016). These results indicate that the actin cytoskeleton may be dispensable for intercellular fusion in embryos, raising the question of whether different types of cell-cell fusion use distinct mechanisms for fusogen recruitment and engagement. Hence, we performed fluorescence time-lapse analysis of cell-cell fusions in RNAi-treated embryos, showing that WASP-Arp2/3-dependent actin polymerization is involved in embryonic hyp7 cell-cell fusions by recruiting EFF-1 to intracellular vesicles. Our results indicate that cell-cell fusions rely on distinct mechanisms at different developmental stages in the formation of a single epithelium syncytium.
The largest C. elegans epithelial hyp7 syncytium contains 139 nuclei, which is formed when embryonic cell-cell fusion merges 23 cells and the remaining 116 cells fuse in larvae (Podbilewicz and White, 1994;Sulston and Horvitz, 1977). To understand whether the actin cytoskeleton plays a general role in C. elegans cell-cell fusion, we examined the dorsal hyp7 cell fusion in Arp2/3 and WASP RNAi embryos. The previous RNAi of the ARX-2 subunit in the Arp2/3 complex did not perturb hyp7 cell fusion, and the animals could survive for >3 days (Patel et al., 2008). Our RNAi treatment caused 100% animal lethality within 2 days (n>200 for each genotype of WASP, Arp2/3 or WAVE), which is suggestive of a more potent depletion. To visualize the cell boundary of two fusing epithelial cells, we constructed a knock-in animal of TagRFP-tagged DLG-1 that is orthologous to the Drosophila Discs large and localizes at apical adherens junctions in all epithelia ( Fig. 1B) (Smurova and Podbilewicz, 2016). We also used GFP or TagRFP-tagged small GTPase MIG-2 to visualize the plasma membrane ( Fig. 2D and Fig. 3A) (Ou and Vale, 2009). We performed fluorescence time-lapse imaging analysis of embryonic cell-cell fusions. A total of 17 dorsal hyp7 precursor cells merged their plasma membranes to form the hyp7 syncytium (Fig. 1A). The loss of DLG-1::TagRFP fluorescence at the border of two hyp7 cells indicates the completion of cell-cell fusion (Fig. 1B).
Considering that EFF-1::GFP forms puncta in the postembryonic seam V.a (anterior daughter cell of the V cell) cells, we examined whether WASP is also involved in the membrane trafficking of EFF-1 in larval seam cells (Yang et al., 2017). Using the same protocol as for EFF-1 quantification in embryos, we found that the WT V.a cells and wsp-1-deficient cells contained 0.14±0.03 and 0.12±0.02 EFF-1::GFP puncta per µm 2 , respectively (mean±s.d., n=16, no statistical significance based on Student's t-test). Although the transmembrane protein EFF-1 must be synthesized in the endoplasmic reticulum and undergo membrane trafficking to arrive on the plasma membrane in both embryonic and larval cells, our results indicate that the underlying regulatory mechanisms appear to be different at distinct developmental stages (Fig. 3D) (Yang et al., 2017).
EFF-1 has distinct localization patterns at different developmental stages in forming a single epithelial syncytium. Although the actin cytoskeleton is generally involved in this process, it facilitates EFF-1 localization in a distinct manner in embryos and larvae. These results indicate that cell-cell fusion may be more complex than is currently appreciated. A previous study revealed that EFF-1 is removed from the plasma membrane by RAB-5 and dynamin-mediated endocytosis in embryos (Smurova and Podbilewicz, 2016). We showed that ARX-2 puncta always localized at the borders of the fusing hyp7 cells (Fig. 2D, n>50 for each marker; Movie 3) and that EFF-1 puncta could occasionally be detected at the border (Yang et al., 2017). Because WASP and Arp2/3 are essential for membrane endocytosis (Firat-Karalar and Welch, 2011), the reduction of EFF-1 puncta in the absence of WASP or Arp2/3 can be explained by defective endocytosis. However, the perturbation of endocytosis led to the hyperfusion phenotype (Smurova and Podbilewicz, 2016) rather than the delay of cell-cell fusion observed in WASP or Arp2/3 RNAi embryos. Given that Arp2/3-based actin nucleation is also required for ER-to-Golgi transport (Campellone et al., 2008), and that EFF-1 puncta colocalize with the Golgi (Smurova and Podbilewicz, 2016), Arp2/3-dependent secretory sorting may facilitate the trafficking of EFF-1 from the ER to the plasma membrane, thereby promoting cell-cell fusion.

C. elegans strains, genetics and DNA manipulations
Strains were maintained on nematode growth medium (NGM) plates seeded with Escherichia coli OP50 following standard protocols at 20°C. CRISPR-Cas9-assisted knock-in animals were generated and analyzed as described previously (Dickinson et al., 2013;Shen et al., 2014). To construct knock-in repair template plasmids, we amplified the 1-1.5 kb upstream and downstream homologous arms from the N2 genomic DNA and inserted them into pPD95.77 with an In-Fusion Advantage PCR Cloning Kit (Clontech, Mountain View, USA). To avoid the cleavage of the homologous repair template by Cas9, synonymous mutations were introduced to the Cas9 target site of the template. The sgRNA plasmid and the knock-in repair template plasmid were co-injected into N2 animals. The knock-in worms were selected and examined by PCR and Sanger sequencing. Transgenic worms were generated by microinjection of DNA plasmids to the germline at 10-50 ng/µl with the co-injection markers pRF4 [rol-6 (su1006)] and Podr-1::dsRed into N2 animals. The primers, plasmids, PCR products and strains are listed in Tables S1-S4.

dsRNA preparation and microinjection
Double-stranded wve-1 RNA was made by in vitro transcription with a T7 RNA RiboMAX Express RNAi System Kit (Promega, Madison, USA). dsRNA was annealed by heating to 90°C for 2 min and cooling by 1°every 8 s until reaching 25°C. dsRNA injections used 1 mg/ml dsRNA in water and were injected to DLG-1::TagRFP knock-in strains.
Live-cell imaging in C. elegans embryo C. elegans eggs in M9 buffer were mounted on 3% agarose pads at 20°C (Chai et al., 2012). Live-cell images were collected with an Axio Observer Z1 microscope (Carl Zeiss MicroImaging, Jena, Germany) equipped with a 100×/1.45 NA objective or a IX83 microscope (Olympus, Southend-on-Sea, UK) equipped with a 150×/1.45 NA oil objective, an EM CCD camera (Andor iXon+DU-897D-C00-#BV-500, Andor Technology, Belfast, UK), and the 488 nm and 568 nm lines of a Sapphire CW CDRH USB Laser System (Coherent, Santa Clara, USA) with a spinning disk confocal scan head (CSU-X1 Spinning Disk Unit, Yokogawa, Kanazawa, Japan). Time-lapse images were acquired with an exposure time of 200 ms every 2 min for imaging the entire cell-cell fusion process. Images were acquired with μManager software (https://www.micro-manager.org/), and processed and quantified with ImageJ software (http://rsbweb.nih.gov/ij/).

Quantification and statistical analysis
To quantify the dorsal hypodermal cell fusions in embryos, we measured the number of the dorsal hyp7 cell borders marked by DLG-1::TagRFP every 10 min from comma stage to twofold stage. EFF-1 puncta (>3 pixel size) density in embryos was measured from the dorsal hypodermal cell 3. To quantify the EFF-1 amount in embryos, we measured the fluorescence intensity of all dorsal hypodermal 7 precursor cells. We used the Student's t-test to determine significant differences in cell-cell fusion between WT and mutants, as indicated in the figure legends.