To investigate the relationship between NRFL-1 and ERM-1, we first examined if NRFL-1 colocalizes with ERM-1. We used CRISPR/Cas9 to engineer an endogenous C-terminal NRFL-1::mCherry fusion, which tags all predicted isoforms. Animals homozygous for the nrfl-1::mCherry knock-in are viable and have a wild-type appearance. We detected expression of NRFL-1 in multiple epithelia including the intestine, excretory canal, pharynx, uterus, and spermatheca (Figures 1B,C). In each of these tissues, NRFL-1::mCherry co-localized with an endogenous ERM-1::GFP fusion protein at the cortex (Figures 1B,C). In the embryo, NRFL-1 localized to the nascent apical domain of intestinal cells, overlapping with ERM-1 (Figure 1B). Confocal super resolution imaging of the intestine in larval stages showed co-localization of NRFL-1 with ERM-1::GFP and YFP::ACT-5 at microvilli, apical to the more intense belt of YFP::ACT-5 at the terminal web (Figure 1C). The observed distribution of NRFL-1::mCherry is consistent with previous observations in C. elegans (Hagiwara et al., 2012), as well as with localization of EBP50 in mammalian epithelial tissues (Ingraffea, 2002; Morales et al., 2004; Kreimann et al., 2007).

We previously showed that ERM-1 and NRFL-1 interact in a yeast two-hybrid assay and in pull-downs from mammalian cultured cells (Koorman et al., 2016). To determine if these proteins interact in a more physiological setting, we used the recently developed split intein-mediated protein ligation (SIMPL) system that relies on protein splicing by split intein domains to detect protein–protein interactions (Yao et al., 2020). We ubiquitously expressed ERM-1 fused to the intein N-terminal fragment (IN) and the V5 epitope, and NRFL-1 fused to the C-terminal fragment (IC) and the FLAG epitope. We observed full splicing of NRFL-1 to ERM-1 by western blot of C. elegans lysates, apparent as a high molecular weight band that stains with both V5 and FLAG antibodies (Figure 2A). In contrast, a negative control pair consisting of IC-tagged NRFL-1 and IN-tagged mKate2 showed only limited splicing of NRFL-1 to mKate2 (Figure 2A). To visualize if splicing occurs in vivo in the intestine, we modified the SIMPL system by including a split mVenus tag. We added the mVenus N-terminal fragment (VN155) to ERM-1::V5-IN and mVenus C-terminal fragment (VC155) to IC-FLAG::NRFL-1, such that upon intein splicing the reconstituted mVenus becomes linked to NRFL-1 (Kodama and Hu, 2010). We readily observed localization of mVenus at the apical domain of intestinal cells, indicating that NRFL-1 and ERM-1 interact in this tissue (Figure 2B).

We next investigated whether NRFL-1 distribution to the apical plasma membrane is dependent on ERM-1, by analyzing NRFL-1::mCherry upon tissue-specific depletion of ERM-1. To deplete ERM-1 in intestinal cells, we introduced an anti-GFP-nanobody::ZIF-1 fusion driven by the intestine-specific elt-2 promoter as an extrachromosomal array in animals expressing endogenous ERM-1::GFP and NRFL-1::mCherry (Wang et al., 2017). Expression of the nanobody::ZIF-1 fusion resulted in variable levels of ERM-1::GFP depletion. The apical levels of ERM-1::GFP and NRFL-1::mCherry showed a linear correlation, indicating that apical recruitment of NRFL-1 in the intestine directly depends on ERM-1 (Figure 2C).

The interaction between mammalian EBP50 and ezrin requires the C-terminal EB domain (Reczek et al., 1997; Reczek and Bretscher, 1998; Finnerty et al., 2004), which is conserved in NRFL-1 (Figures 1A; Supplementary Figure S1A). To determine if the NRFL-1 EB domain is required for the interaction with ERM-1, we repeated the SIMPL-mVenus experiment using an NRFL-1(ΔEB) mutant that lacks the C-terminal 28 amino acids of NRFL-1. Compared to wild-type NRFL-1, we observed only residual apical localization of mVenus in intestinal cells, indicating that the interaction of NRFL-1 with ERM-1 depends on the presence of the EB domain (Figure 2B).

To determine if the EB domain is necessary for the apical localization of NRFL-1, we used CRISPR/Cas9 to engineer the 28 aa EB deletion in the nrfl-1::mCherry strain. The resulting nrfl-1(Δeb)::mCherry animals are homozygous viable, consistent with the lack of severe defects in previously described nrfl-1 mutants (Hagiwara et al., 2012; Na et al., 2017). We detected a dramatic reduction in apical levels of NRFL-1(ΔEB)::mCherry in intestinal cells when compared with NRFL-1::mCherry (Figure 2D). NRFL-1(ΔEB)::mCherry also failed to localize at the cortex in the uterus and spermatheca, while apical levels in the excretory canal were reduced (Figure 2E). These results indicate that apical recruitment of NRFL-1 is mediated by the EB domain. However, the presence of some residual apical NRFL-1(ΔEB)::mCherry in the intestine and excretory canal suggests the existence of alternative membrane-targeting mechanisms. Collectively, our results show that the interaction between ERM and NHERF proteins is conserved in C. elegans, and that the localization of NRFL-1 is largely mediated by its interaction with ERM-1.

We next wanted to investigate the effects of loss of NRFL-1 on intestinal lumen formation. Previous studies using partial deletion alleles of nrfl-1 indicated that loss of NRFL-1 alone does not cause defects in the formation of the intestine (Hagiwara et al., 2012; Na et al., 2017). To rule out the possibility that the lack of severe defects is due to the production of truncated NRFL-1 proteins, we used CRISPR/Cas9 genome engineering to generate the nrfl-1(mib59) deletion allele. This allele lacks almost the entire nrfl-1 locus and additionally causes a frameshift in the first exon of the long isoforms (Figure 3A). Hence, we refer to mib59 as nrfl-1(null). The mib59 deletion also removes a candidate non-coding RNA and overlapping 21U-RNA located in the large 3^rd exon of nrfl-1a. Animals homozygous for the nrfl-1(null) allele are viable, have a healthy appearance, and normal brood sizes, confirming that NRFL-1 is not essential for C. elegans development (Figure 3B,C).

One of the possible reasons for the lack of a severe intestinal phenotype in nrfl-1(null) animals is that NRFL-1 may only mediate part of the functions of ERM-1 in the intestine. To investigate this possibility, we made use of the non-phosphorylatable erm-1[T544A] and phosphomimetic erm-1[T544D] alleles we generated previously (Ramalho et al., 2020). Both mutants cause a delay in the apical recruitment of ERM-1 and actin during embryogenesis, and the appearance of constrictions along the course of the lumen that only occasionally persist to the L1 stage. In contrast to erm-1 RNAi or strong loss-of-function alleles, however, these animals are viable. Thus, erm-1[T544A] and erm-1[T544D] represent partial loss-of-function alleles that may act as a sensitized background to reveal the contribution of NRFL-1 to ERM-1 functioning. We therefore generated double mutants that carry the nrfl-1(null) allele and either of the erm-1[T544A] or erm-1[T544D] alleles. As a first indicator of synthetic defects, we examined the double mutant strains for embryonic lethality or an increase in the mild brood size defect observed in erm-1[T544A] and erm-1[T544D] mutants. We did not observe strong embryonic lethality in any mutant combination (<5%, Figure 3B). However, combining nrfl-1(null) with either erm-1 phosphorylation mutant resulted in a strongly reduced brood size (Figure 3C). In addition, many larvae in the double mutant combination had a sick appearance and developed slowly. Nevertheless, both double mutants can be maintained as homozygotes, unlike strong erm-1 loss of function mutants.

We next examined the formation of the intestinal lumen and actin distribution using YFP::ACT-5 as a marker. We did not detect any defects in apical enrichment of ACT-5 or intestinal morphology in nrfl-1(null) embryos and larvae (Figures 4A–C). Combining the erm-1[T544A] and erm-1[T544D] alleles with the nrfl-1(null) allele significantly increased the frequency of intestinal constrictions and their persistence until larval development (Figures 4A,C). Intestines of early larval nrfl-1(null); erm-1[T544A] and nrfl-1(null); erm-1[T544D] animals were characterized by a cystic appearance and multiple constrictions that block intestinal flow as seen in feeding assays with fluorescent membrane-impermeable dextran (Supplementary Figure S2A). In surviving L2 or older animals, we only observed morphological defects but no lumen discontinuities, indicating that the early larval arrest in double mutants is due to a block of flow of food through the intestine (Supplementary Figure S2B). In addition to the increase in intestinal constrictions, we also observe that loss of nrfl-1 caused a further decrease in the apical levels of YFP::ACT-5 in erm-1[T544A] and erm-1[T544D] mutant animals (Figure 4B).

Finally, as the EB domain is essential for the apical localization of NRFL-1 and its interaction with ERM-1, we determined if loss of the EB domain results in similar synergistic phenotypes with the ERM-1 phosphorylation mutants as complete loss of NRFL-1. We used CRISPR/Cas9 genome engineering to generate a second nrfl-1(Δeb) allele, also removing the final 28 aa but lacking the mCherry tag used above (Figure 3A). Similar to our observations for the mCherry-tagged variant, homozygous nrfl-1(Δeb) mutants are viable and show no significant defects in brood size, intestinal development, or apical ACT-5 enrichment (Figures 3B,C; Figures 4A–C). However, when combined with erm-1[T544A] or erm-1[T544D], the resulting double mutants showed similar defects in viability, growth, brood size, intestinal development, and ACT-5 enrichment as observed using the nrfl-1(null) allele (Figures 3B,C; Figures 4A–C). Thus, the nrfl-1(Δeb) allele behaves like a null allele of nrfl-1. Taken together, our data show that NRFL-1 and ERM-1 function together in promoting lumen formation in the C. elegans intestine, and the binding to ERM-1 is essential for the functioning of NRFL-1 in the intestine.

NRFL-1 could function together with ERM-1 in at least two ways. It could act as a scaffold protein that is required for ERM-1 to organize protein complexes at the membrane, or it could regulate the activity of ERM-1 itself. To distinguish between these possibilities, we investigated whether loss of NRFL-1 affects the distribution, mobility, or T544 phosphorylation status of ERM-1. We first analyzed the distribution of ERM-1::GFP in larval nrfl-1(null) mutants. We did not detect any change in ERM-1::GFP subcellular localization or levels at the apical membrane in the intestine (Figure 5A). Moreover, FRAP analysis demonstrated that the mobility of ERM-1::GFP at the apical intestinal membrane was not significantly altered in nrfl-1(null) larvae (Figure 5B). We next investigated whether NRFL-1 regulates ERM-1 C-terminal phosphorylation by staining nrfl-1(null) mutants with an antibody specific for the C-terminal phosphorylated form of ERM proteins (pERM). The residues used to raise this antibody are fully conserved between mammals and C. elegans (Ramalho et al., 2020). Nevertheless, we first confirmed the specificity of the antibody for T544 phosphorylated ERM-1 by immunostaining of ERM-1[T544A] mutant animals. We readily detected pERM staining of the intestinal lumen in wild-type larvae, while no staining was observed in ERM-1[T544A] animals (Supplementary Figure S3A). Moreover, treatment of embryos with a phosphatase abolished staining with the pERM antibody (Supplementary Figure S3B). Thus, the pERM antibody is specific for T544 phosphorylated ERM-1. We then stained nrfl-1(+) and nrfl-1(null) animals with the pERM antibody. In both backgrounds, the pERM antibody stained the lumen of the intestine, indicating that loss of nrfl-1 does not significantly alter the phosphorylation status of the C-terminal regulatory threonine of ERM-1 (Figure 5C). Taken together, our results show that NRFL-1 does not regulate the distribution, dynamics, or phosphorylation of ERM-1, and therefore does not seem to directly regulate ERM-1.

C. elegans strains were cultured under standard conditions (Brenner, 1974). Only hermaphrodites were used, and all experiments were performed with animals grown at 15 °C or 20 °C on standard Nematode Growth Medium (NGM) agar plates seeded with OP50 Escherichia coli. Table 1 contains a list of all the strains used.

Bait and prey SIMPL constructs were generated using the SapI-based cloning strategy, as previously described (Yao et al., 2020). For the conventional SIMPL system, previously described intein inserts were used (Yao et al., 2020). For the SIMPL-mVenus system, the InteinC-3xFLAG-VC155 and VN155-HA-V5-InteinN inserts were codon-optimized for C. elegans, flanked by SapI sites and ordered as gBlocks (IDT). Primers containing the appropriate SapI overhangs were used to amplify erm-1, nrfl-1 and nrfl-1(Δeb) from a cDNA library, InteinC-3xFLAG-VC155 from the ordered gBlock and mKate from pDD375 (Addgene #91825). All gBlocks and PCR products were blunt-end cloned into the plasmid pHSG298. Bait or prey, intein, the rps-0 promoter and the unc-54 3’ UTR fragments were combined and inserted into the pMLS257 plasmid (Addgene #73716) using the SapTrap assembly method (Schwartz and Jorgensen, 2016; Yao et al., 2020). Finally for the SIMPL-mVenus system, an mKate2 sequence was integrated into the newly generated Prps-0::erm-1::VN155-HA-V5-InteinN::unc-54 plasmid. The mKate2 sequence and the Prps-0::erm-1:VN155-HA-V5-InteinN::unc-54 plasmid were amplified using primers with the appropriate overhangs to incorporate the mKate2 into the plasmid between the promotor and erm-1 coding sequence using Gibson Assembly (GA). Constructs were verified by Sanger sequencing before injection (Macrogen Europe). Plasmids used for injection were purified using the PureLink HQ Mini Plasmid DNA Purification Kit (Thermo Fisher) using the extra wash step and buffer recommended for endA + strains. Final plasmid sequences are available in Genbank format in Supplementary File S1.

Transgenic animals expressing bait and prey constructs were generated by microinjection in the gonads of young adult N2 animals using an inverted microinjection setup (Eppendorf) with 20 ng/μL of bait and prey plasmids, as well as the pDD382 plasmid (Addgene #91830) containing a visible dominant Rol marker and an hygromycin selection cassette. The DNA mix was spun at max speed on a tabletop centrifuge for 15 min prior to injection. Injected animals were incubated for 2–3 days at 20°C before addition of hygromycin B (250 μg/ml) to the plates. After 1–2 days, surviving Rol animals were singled, allowed to develop, and F2 progeny was screened for successful transmission of the transgenic extrachromosomal array. Multiple lines with successful transmission were saved and used for analysis.

Animals were grown on NGM plates supplemented with hygromycin B (250 μg/ml) until plates were full, washed off with M9 buffer (0.22 M KH₂PO, 0.42 M Na₂HPO₄, 0.85 M NaCl, 0.001 M MgSO₄), washed three times with M9 buffer, and incubated at room temperature (RT) for 20 min. Samples were then pelleted and resuspended in 100–200 µL of lysis buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.5% IGEPAL CA-630 (Sigma-Aldrich), 1 tablet/50 ml cOmplete protease inhibitor cocktail (Sigma-Aldrich)), and sonicated with a Diagenode BioRupter Plus for 10 min with the high setting and on/off cycles of 30 s in a 4 °C water bath. The lysates were spun at max speed for 15 min, an equal volume of 2 × SDS buffer (100 mM Tris-HCl, 4% SDS, 0.2% bromophenol blue, 20% glycerol, and 10% β-mercaptoethanol) was added, and boiled 10 min. Depending on the experiment, 5–12 µL of protein lysate was loaded into pre-cast protein gels (4–12% Bolt Bis Tris Plus, ThermoFisher) together with 10 µL of the molecular marker (PageRuler prestained, ThermoFisher). Gels were run for 30–45 min at 200 V in NuPAGE MOPS SDS Running buffer (ThermoFisher), and transferred onto a PVDF membrane (Immobilon-P 0.45 µm, Millipore) at 4°C and 30 V overnight in Bolt transfer buffer (Thermo Fisher). For staining, membranes were rinsed in TBST (50 mM Tris-Cl, 150 mM NaCl, 0.1% Tween-20), blocked with 4% milk in TBST for 1 h at RT, and incubated with primary antibodies in milk for 1 h at RT. Membranes were washed three times for 10 min in TBST, incubated with secondary antibodies in milk for 1 h at RT, and washed again three times for 10 min in TBST before exposure using ECL (SignalFire Plus, Cell signaling). The following antibodies and concentrations were used: rabbit anti-V5, 1:1000 (Cell Signaling #13202); mouse anti-FLAG, 1:10000 (Sigma #F1804); goat anti-Rabbit and donkey anti-mouse HRP conjugates, 1:5000.

The nrfl-1::mCherry, nrfl-1(Δeb) and nrfl-1(Δeb)::mCherry strains were engineered by homology-directed repair of CRISPR/Cas9-induced DNA double-strand breaks (DSBs), while the nrfl-1(null) deletion was generated by imprecise repair of CRISPR/Cas9-induced DSBs. Delivery of components for CRISPR/Cas9 editing was done by microinjection in the gonads of young adult animals of different genetic backgrounds: nrfl-1(Δeb) and nrfl-1(mib59) were generated in an N2 background; nrfl-1::mCherry was generated in BOX273 background, and nrfl-1(Δeb)::mCherry in a BOX422 background. All sequences of the oligonucleotides and crRNAs used (synthesized by IDT) are listed in Table 2.

For the nrfl-1::mCherry, two plasmid-based sgRNAs were used, generated by ligation of annealed oligo pairs into the pU6::sgRNA expression vector pJJR50 (Addgene #75026) as previously described (Waaijers et al., 2016). To generate the nrfl-1::mCherry repair template we created a custom SEC vector, pJJR83 (Addgene #75028), by replacing a fragment of pDD282 (Addgene #66823) containing the GFP sequence with a similar fragment containing a codon optimized mCherry sequence with synthetic introns using the flanking Bsu36I and BglII restriction sites. Homology arms of about ±750 bp, flanking the DSB site, were amplified from genomic DNA and introduced into pJJR83 as previously described (Dickinson et al., 2015). The sgRNA (100 ng/μL) and SEC repair template (20 ng/μL) plasmids combined with Peft-3::Cas9 (60 ng/μL; Addgene #46168) and Pmyo-2::mCherry co-injection marker (2.5 ng/μL; pCFJ90, Addgene #19327) were micro-injected in the gonad of young adults. Two injected animals were pooled per plate, incubated for 3 days at 20°C, 500 μL of 5 mg/ml hygromycin was added per plate, and non-transgenic Rol animals were selected after 4–5 days. These selected animals were lysed and genotyped with primers flanking the homology arms and confirmed by Sanger sequencing. To eliminate the SEC selection cassette L1 progeny of homozygous Rol animals was heat shocked in a water-bath at 34°C for 1 h.

To generate the nrfl-1(null) deletion allele a mix containing Peft-3::Cas9 (Addgene #46168; 50 ng/μL), two pairs of sgRNA plasmids targeting the 5′ or 3′ ends of the nrfl-1 open reading frame (75 ng/μL each), and a dpy-10 sgRNA plasmid (50 ng/μL) for co-CRISPR selection (Arribere et al., 2014) were micro-injected in the gonad of young adults. To select for deletions, injected animals were transferred to individual plates, incubated for 3–4 days at 20°C, and 96 non-transgenic F1 animals (wild-type, Dpy, or Rol) from 2–3 plates containing high numbers of Dpy and Rol animals were selected and transferred to individual plates. After laying eggs, F1 animals were lysed and genotyped with primers flanking the nrfl-1 ORF. In all cases, deletions were confirmed by Sanger sequencing. Sanger sequencing was also used to determine the precise molecular lesion in selected animals. The nrfl-1(null) allele used in this paper, nrfl-1(mib59), consists of a 9 bp deletion starting 1 bp before the initial base of the start codon of long nrfl-1 isoforms (a, c, d, h, j), and a second 11,537 bp deletion spanning part of the third exon (791 bp from start of nrfl-1a) until the downstream intergenic region, which includes the entire ORFs of the small nrfl-1 isoforms (left flank 5′- atg​ctt​gtg​atc​tct​gaa​gaa​gga​g, right flank 5′ aat​atc​acg​aac​aac​ttc​tag​gag​c). The mib59 allele also deleted an ncRNA (C01F6.16) and three piRNAs (C01F6.10, F32B2.25, and F23B2.28) located within nrfl-1 introns.

The NRFL-1 EB domain deletions were generated using the Alt-R CRISPR-Cas9 system (IDT). A single-stranded oligodeoxynucleotide with about 35 bp homology arms was used as a repair template to fuse the flanks of a deletion spanning nucleotides 1390–1473 of nrfl-1h, as previously described (Dokshin et al., 2018). A mix of 250 ng/μL Cas9 protein, 2 μM repair template, 4.5 μM each nrfl-1 crRNAs, 10 μM tracrRNA, as well as 1 μM dpy-10 crRNA and ssODN repair for co-CRISPR selection (Arribere et al., 2014) was micro-injected into the gonads of young adults. Animals were selected as described above for the nrfl-1(null) allele and genotyped using two primers flanking the deletion.

Imaging of C. elegans was done by mounting embryos or larvae on a 5% agarose pad in 20 mM Tetramisole solution in M9 to induce paralysis. Spinning disk confocal imaging was performed using a Nikon Eclipse Ti-U manual microscope equipped with a Yokogawa CSU-X1 spinning disk using a 60× 1.4 NA objective, 488 and 561 nm lasers, Semrock 488 long-pass, 525/30 (green), 617/73 (red) & 512/630 (dual) emission filters, 600 Texas Red (EX540-580/DM595/BA600-660) filter blocks, and Andor iXON DU-885 camera. Imaging for FRAP and immunohistochemistry experiments was performed on a Nikon Eclipse-Ti with Perfect Focus System microscope equipped with a Yokogawa CSU-X1-A1 spinning disk using 60× and 100× 1.4 NA objectives, Chroma ET-DAPI (49000), ET-GFP (49002), ET-mCherry (49008) emission filters, 355 nm, 488 nm, 491 nm, and 561 nm lasers, and a Photometrics Evolve 512 EMCCD camera. Targeted photobleaching was done using an ILas system (Roper Scientific France/PICT-IBiSA, Institut Curie). Spinning disk images were acquired using MetaMorph Microscopy Automation and Image Analysis Software. All stacks along the z-axis were obtained at 0.25 μm intervals. Super resolution images of the microvilli were obtained using a Zeiss AxioObserver 7 SP microscope with Definite Focus 2 operated by Zeiss ZEN software with an Airyscan 32-channel GaAsP-PMT area detector using a 100× 1.46 NA objective, and Laser Argon Multiline and 561 nm lasers. Maximum intensity Z projections were done in ImageJ (Fiji) software (Schindelin et al., 2012; Rueden et al., 2017). For quantifications, the same laser power and exposure times were used within experiments. Image scales were calibrated for each microscope using a micrometer slide. For display in figures, level adjustments, false coloring, and image overlays were done in Adobe Photoshop. Image rotation, cropping, and panel assembly were done in Adobe Illustrator. All edits were done non-destructively using adjustment layers and clipping masks, and images were kept in their original capture bit depth until final export from Illustrator for publication.

Quantitative analysis of spinning disk images was done in Fiji. All values were corrected for background levels by subtracting the average of three regions within the field of view that did not contain any animals. For quantification of apical protein levels, measurements were done in intestinal cells forming int2 through int6, and where the opposing apical membranes could be clearly seen as two lines. Levels were obtained by averaging the peak values of intensity profiles from three 25 px-wide (10 px-wide for the SIMPL-mVenus system) line scans perpendicular to the membrane per animal. For YFP::ACT-5, which is expressed from a transgene with variable expression levels, we express apical enrichment as the ratio of apical/cytoplasmic. Cytoplasmic levels were measured by averaging three regions within the cytoplasm of intestinal cells. Intensity distribution profiles to analyze co-distribution of NRFL-1 with ERM-1 and ACT-5 were obtained by taking three 25 px-wide line scans perpendicular to the apical membrane in each animal. Before averaging these three values, they were aligned and normalized to the peak value. Measurements of multiple animals were again aligned based on the peak value. All presented graphs were made using GraphPad Prism and Adobe Illustrator.

For protein degradation using the anti-GFP-nanobody::ZIF-1 approach (Wang et al., 2017), gonads of young adult BOX428 animals were microinjected with 30 ng/μL Pelt-2::α-GFP-NB::ZIF-1 and 2.5 ng/μL Pmyo-2::GFP (#Addgene 26347) as a co-injection marker. Transgenic F1 animals were transferred to individual plates, F2 progeny was screened for successful transmission of the extrachromosomal array and imaged using spinning disk microscopy.

L4 animals were put on individual plates at 20°C and transferred to a new plate daily until they died. After the parent was removed from a plate, hatched animals and the unhatched eggs were counted 2–4 days later. The number of animals and unhatched eggs combined constitutes the total progeny size. The graph presented was made using GraphPad Prism and Adobe Illustrator.

Mixed stage populations were collected in M9 and washed two times in M9. Animals were then pelleted, concentrated, resuspended in 1 mg/ml Texas Red-dextran 40,000 MW (Thermofisher D1829) in egg buffer (118 mM NaCl, 48 mM KCl, 2 mM MgCl2, 2 mM CaCl2, 25 mM HEPES pH 7.3), and incubated for 60 min on a shaker at 500 rpm. The dye in solution was removed by washing the samples with M9 two times. Animals were paralyzed in 10 mM Tetramisole, transferred to an agarose pad on a glass slide, and imaged using spinning disk microscopy.

For FRAP assays, laser power was adjusted in each experiment to avoid complete photobleaching of the selected area, as the time scale of experiments prevented assessment of photo-induced damage. Photobleaching was performed on a circular region with a diameter of 30 or 40 px at the cortex, and images were taken just before bleaching, directly after, after 15 min, and after 45 min. These images were analyzed using ImageJ. The size of the area for FRAP analysis was defined by the full width at half maximum of an intensity plot across the bleached region. For each time point, the mean intensity value within the bleached region was determined, and the background, defined as the mean intensity of a non-bleached region outside the animal, was subtracted. The mean intensities within the bleached region were corrected for acquisition photobleaching per frame using the background-subtracted mean intensity of a similar non-bleached region at the cortex, which was normalized to the corresponding pre-bleach mean intensity. FRAP recovery was calculated as the change in corrected intensity values within the bleached region from the first image after bleaching normalized to the mean intensity just before bleaching.

For the staining of larval stages, embryos were obtained from gravid adults by bleaching and allowed to hatch and develop on plates at 15°C for 24 h. Animals were collected from plates and washed three times with M9 and once with MQ H₂O before being transferred to poly-L-lysine-coated frosted slides. For the staining of embryos, embryos were obtained from gravid adults by dissection in MQ H₂O on poly-L-lysine-coated frosted slides and allowed to develop at RT for 4 h. A coverslip (Carl Roth, #1) was lowered on top of larvae/embryos, followed by freezing in liquid nitrogen and snapping off of the coverslip. Fixation was performed in formaldehyde solution with phosphatase inhibitors (3,7% formaldehyde (Sigma-Aldrich), 250 µM EDTA and 50 mM NaF in PBS (1,35 M NaCl, 27 mM KCl, 100 mM Na₂HPO₄, 18 mM KH₂PO₄)) at RT for 10 min. Samples were rinsed in PBS, permeabilized (PBS + 0,5% triton X-100 (Sigma-Aldrich)) for 30 min, washed four times in wash buffer (0,1% Triton X-100, 250 µM EDTA and 50 mM NaF in PBS) for 10 min each and then blocked (1% bovine serum albumin (Sigma-Aldrich) and 10% goat serum (Sigma-Aldrich)) for 1 h at RT. For the staining with protein phosphatase treatment, samples were treated with Lambda Protein phosphatase (NEB) for 30 min at 30°C followed with an additional four times washing step before they were blocked. Primary antibodies (anti-phospho-ezrin (Thr567)/radixin (Thr564)/moesin (Thr558) (48G2) rabbit mAb #3726 (Cell Signaling Technologies) 1:200 and mouse anti-DLG (Hybridoma bank) 1:50) in blocking solution were applied overnight at 4°C. Samples were then washed four times in wash buffer for 10 min each and stained with secondary antibodies (Alexa-Fluor 488 goat anti-rabbit and Alexa-Fluor 568 goat anti-mouse (Life Technologies, A-11008 and A11004), both 1:500) in blocking solution for 1 hour at RT. Samples were then washed four times in wash buffer and once in PBS for 10 min each and finally mounted with Prolong Gold Antifade with DAPI (Thermofisher) under a coverslip and sealed with nail polish.

All statistical analyses were performed using GraphPad Prism 8. For population comparisons, a D’Agostino and Pearson test of normality was first performed to determine if the data was sampled from a Gaussian distribution. For data drawn from a Gaussian distribution, comparisons between two populations were done using an unpaired t-test, with Welch’s correction if the SDs of the populations differed significantly, and comparisons between >2 populations were done using a one-way ANOVA, or a Welch’s ANOVA if the SDs of the populations differed significantly. For data not drawn from a Gaussian distribution, a non-parametric test was used (Mann-Whitney for 2 populations and Kruskal-Wallis for >2 populations). ANOVA and non-parametric tests were followed up with multiple comparison tests of significance (Dunnett’s, Tukey’s, Dunnett’s T3 or Dunn’s). Tests of significance used and sample sizes are indicated in the figure legends. No statistical method was used to pre-determine sample sizes. No samples or animals were excluded from analysis. The experiments were not randomized, and the investigators were not blinded to allocation during experiments and outcome assessment.