Reciprocal E-cadherin signaling aligns apical surfaces between neighboring epithelial tissues to complete the C. elegans digestive tract

Summary

Epithelial cells form barriers with specialized apical membranes facing the external environment or internal lumens, yet how interconnected epithelial tubular systems form from cystic epithelial primordia is poorly studied. At the boundaries of cystic epithelial primordia, apical surfaces must correctly align to yield functional connections within and between organs. Here, we use the digestive tract of the developing C. elegans embryo to define a crucial two-cell tissue, the rectal valve (vir) cells, that mediate the connection and apical alignment between the posterior intestine and the rectum. Vir cells migrate while attached to the rectal primordium to create an essential connection with the intestinal primordium and complete the digestive tract. Vir contact with the intestine, through E-cadherin signaling, is required for the intestinal apical surface to undergo a transition from a cyst with an inaccessible apical surface to an open tube competent to connect with neighboring tissues. Ablation of vir cell progenitor resulted in a lethal disconnection between the intestinal and rectal apical surfaces due to failure of this first step in apical accessibility. To resolve the final gap in apical continuity between the intestine and rectum, vir cells develop bipolar apical surfaces, each facing the neighboring tissue. The intestine-facing apical puncta in vir cells require intestinal E-cadherin signaling to form and accumulate as well as to correctly partition apical proteins within vir cells. Together these results establish E-cadherin-based signaling as the crucial factor aligning apical surfaces across primordia as epithelial cells from different germ layers form a vital connection.

eTOC

Animals are often encased by a continuous epithelium, presenting a significant morphogenetic challenge where tissues of different origins must connect during development. Cote et al show how reciprocal E-cadherin-mediated signaling between the intestine and a two-celled bipolar tissue, the rectal valve, completes a functional digestive tract during C. elegans embryogenesis.

Introduction

The most common body plan in the animal kingdom is an ‘epithelial toroid’ where epithelial cells from endoderm and ectoderm connect to form a continuous casing from exterior of the body through to a central digestive tube^1,2. These epithelial cells have specialized non-adherent apical membranes that face internal lumens or the external environment and are bordered by cell-cell junctions^3–8. In culture models, epithelia readily build individual closed-off cysts, i.e. cells fully surrounding a central lumen, recapitulating aspects of lumen formation within developing primordia. Yet, how enclosed cysts within a complicated multi-tissue environment connect to form functional interconnected tubular systems is poorly understood.

At the cyst-cyst interface between epithelia primordia, cells that may have different developmental origins, polarization mechanisms, and functions must nonetheless align to create a single continuous apical surface. For example, in kidney morphogenesis, the collecting duct and renal vesicle epithelial populations undergo apical polarization at different times, form spatially separate lumens, yet must fuse to form a functional nephron⁹. Similarly in sensory, reproductive, excretory, circulatory, and digestive systems, epithelial cells from both within and between different tissues must form continuous apical surfaces^10–15. Ectopic or inappropriate connections between different tissues can lead to fistulas while absence of correct apical alignment between fusing epithelial structures can lead to congenital atresias during ear, esophagus, urogenital, and bile duct development^16–20.

Although decades of careful cell biology have revealed how epithelial cells within one tissue undergo apical polarization^1,3,7, how the cells at the boundaries between neighboring tissues align their apical surfaces remains understudied due to the need to generate multiple sets of tissue-specific tools. Here, we use the development of the C. elegans digestive tract to investigate the alignment of apical surfaces across tissues, focusing on the final connection between the intestine and rectum that completes the epithelial toroid (Figure 1A–C).

Figure 1: Rectal valve cells (vir) are a two-cell tissue that mediates the intestine-rectal connection.

(A) Live image of 1.5-fold stage embryo expressing endogenously-tagged PKC-3 (black, ‘Apical’ throughout) and digestive tract plasma membrane (pha-4p::PM recolored by tissue^†). valve-intestinal-rectal (vir, orange throughout) cells localize within final apical discontinuity between the rectum and intestine (teal throughout). Orange box, inset location in all figures unless noted, depicting orientation (rectum on left).

(B) Tracing of continuous larval lumen (black) from serial section EM²⁹.

(C) Morphogenetic timeline of posterior intestine and vir. mpf: minutes post fertilization with timing relative to vir birth used in all timelapses. Vertical lines: cell divisions. E8-E16-E20: intestinal ‘E’ cell number.

(D, E) Larval intestine-rectal connection (PKC-3, magenta; EPS-8A, ceh-27p::HistoneH1, green) after embryonic progenitor laser cell ablation. *, intestinal cyst. All figures unless noted: live confocal image of C. elegans embryo or larvae expressing endogenous fluorescently-tagged protein(s) (‘max’: maximum intensity projection through tissue; ‘slice’: single z-slice). Single channel images boxed in corresponding color. Scale bars, 5 μm.

See also Figure S1.

Results

Rectal valve cells (vir) mediate the intestinal-rectal connection

To determine how the continuous apical surface of the epithelial toroid develops, we explored the endogenous localization of the fluorescently-tagged apical protein aPKC/PKC-3. We found a continuous line of apical proteins from the skin through the pharynx to the intestine, but observed a break in the continuity of the apical surface between the intestinal primordium and the invaginating rectal lumen at the posterior of the embryo indicating that this surface was the final connection completing the epithelial toroid (Figure 1A colored by tissue, n = 303/305 1.5-fold embryos, Figure S1A’). At this 1.5-fold stage when the embryo is 1.5 times its initial length and folded inside the eggshell, both the intestine and rectum were polarized with apical and junctional proteins concentrated towards the lumen^4,21–27. However, these apical surfaces were separated by a pair of cells sitting at their interface that corresponded to the rectal valve cells²⁸ (‘vir’, valve-intestinal-rectal: virR and virL, orange), resulting in a gap in apical continuity (Figure 1C, Figure S1A–E). In serial section electron microscopy (EM) of larvae²⁹ that shows the final form of the digestive tract, the ~2um long interface between vir cells contains a thin (55nm wide) and electron-dense sieve between the intestinal and rectal lumens (Figure 1B, Figure S1F) that is opened and closed by the surrounding muscle during waste expulsion³⁰. Ablation of the vir progenitor before vir-intestinal contact resulted in a disconnection between the intestinal and the rectal lumens²⁸ (Figure 1D–E), although vir cells were dispensable for rectal invagination (Figure S1G–H). The resulting ‘blind gut’ caused arrest at the L1 larval stage, consistent with genetic evidence that rectal deformities are associated with lethality^31–34, presumably due to the inability of the worms to expel waste. Therefore, vir cells provide a specific and essential luminal connection that completes the digestive tract.

To investigate the morphogenesis of the rectal valve, we first observed that the two vir cells arise from a single progenitor that is lineally and spatially distinct from the intestinal progenitor (Figure 2A, Figure S2, Video S1). The progenitors of vir and the 9 other rectal cells are born on the ventral surface and all arise from the AB-lineage²⁸. At this stage of development, the E blastomere-derived intestine is composed of 16 cells (‘E16’) that are named according to which of the nine intestinal rings (int1-int9) they give rise to³⁵. Live imaging revealed that the vir progenitor gastrulated with its non-rectal neighbors to internalize³⁶ (Figure S2C), divided before the remaining rectal cells invaginated, and was initially separated from the intestine by muscle and germ cells³⁷ and not a laminin-rich basement membrane (Figure 2A–B). The vir cells contacted the intestinal primordium (Figure 2A, Figure S2B, Video S1) after the intestine had polarized³⁸ (Figure S2D) and well before the final intestinal divisions (‘E16-E20’, Figure S2G). Whole-embryo cell membrane segmentation³⁷ showed that virL and virR stably contacted the posterior-most intestinal cells (int8L and in8R) ~15 min after vir cell birth (Figure S2B, Video S1) during the intercalation of int5 cells within the intestinal primordium³⁵ (Figure S2E), although transient interactions between the intestine and vir progenitor or the posterior virL occurred in 3/8 embryos. After prolonged intestinal contact, the more anterior virR cell then rotated to form a bilaterally symmetric interaction with the posterior end of the intestine (Figure S2F). At the 1.5-fold stage, about 2 hours after the initial contact between the vir cells and the posterior intestine, the gap in apical continuity between the intestine and rectum (Figure 1A) started to close, with the intestinal apical surface appearing to extend towards and join with the apical surface of the rectum during embryo elongation (Figure 2C–F, Figure S3, Video S2 detailing apical highlight).

Figure 2: vir cells migrate to the intestine and are located at the site of apical continuity between the intestine and rectum, independent of a basement membrane breach.

(A) Timelapse of intestinal (end-1p::PM, magenta) and virR/L (nhr-67p::PM, magenta; FOS-1A, green nuclei) membranes and basement membrane (LAM-2/laminin, green) of embryo in lateral view. ‘G’ indicates primordial germ cells.

(B) Valve cells do not breach a laminin+ membrane (representative of n=28 time courses across 4 strains). Ventral view of 3D rendering of LAM-2::mNeonGreen shown with outlines of vir cells drawn from nhr-67p::PM outlines of GFP::FOS-1A+ cells. Box indicates location of detailed ventral view time course on right.

(C-F) Live imaging of merging apical surfaces (PKC-3, magenta, highlighted in D (see Methods)) during embryonic elongation within the digestive tract (pha-4p::PM, green) with vir marker (nhr-67p::H1, magenta) or with apical junctions (E-cadherin/HMR-1, green). Cartoons in D aligned to rectum.

All figures are live confocal images of C. elegans embryo expressing endogenous fluorescently-tagged protein(s) (‘max’: maximum intensity projection through tissue; ‘slice’: single z-slice; ‘3D’: 3-dimensional napari rendering). Single channel images boxed in corresponding color. Scale bars, 5 μm.

See also Figure S2, Figure S3, Video S1, and Video S2.

Specialized cells often breach a laminin-rich basement membrane to facilitate the joining of tissues, as in the C. elegans vulva, mammalian mouth, and kidney^39–42. However, we found no laminin between vir cells and the posterior end of the intestine, consistent with prior descriptions of extracellular membrane proteins throughout embryogenesis⁴³, and instead laminin surrounded the digestive tract after the vir cells made stable contact with the intestine (Figure 2A–B). Therefore, the vir cells form a specialized tissue that bridges between the internal intestinal lumen and the externalfacing rectal lumen through cell-cell contact.

The rectal valve cells are required for a cyst-to-tube transition in the intestine

Given that larvae after vir precursor ablation showed an abrupt end to the intestinal lumen (Figure 1D, Figure S1H), we next explored how the apical surface of the intestine normally joins with that of the rectum during morphogenesis. Imaging from the ventral surface during the early E16 stage showed that the intestinal apical surface was present at the time when the vir cells made contact but did not reach the end of int8L/R (Figure S2D); the posterior-most intestinal cells formed a closed off cyst with an apical surface that was inaccessible to other cells. About 10 minutes after the valve cells contacted the intestine, the intestinal apical surface appeared to extend and erode the cell-cell interface between the posterior int8L and int8R cells, converting the cyst into an accessible tube by redistributing apical proteins (Figure 3A–B, Figure S4A–C, Video S3, see Methods). To test whether interaction with vir cells was required for the intestine to undergo this cyst-to-tube transition, we examined embryos at late bean stage, the time when the cyst-to-tube transition is complete, after laser cell ablation of the vir progenitor. Control ablations of cells neighboring the vir progenitor at the ventral midline resulted in a normal cyst-to-tube transition, indicating that thermal damage did not delay or block this apical extension within the intestinal cells (Figure 3C–D, Figure S4D–E). In contrast, valve progenitor ablation blocked the cyst-to-tube transition, indicating that the ‘blind gut’ visible in larvae (Figure 1D–E, Figure S1H) was a result of a failure of the first step in joining the intestine lumen to the rectum: creating an accessible apical surface at the end of the intestine.

Figure 3: Rectal valve cells induce a cyst-to-tube transition in the intestine mediated by E-cadherin.

(A) Ventrolateral 3D renderings of apical PKC-3 time course along the int8L/R interface (light green) with 3D cell reconstructions from pha-4p::PM. (A’) Appearance of small puncta (orange arrowhead) at the valve/intestinal contact. scale bar, 2 μm.

(B) Percentage coverage of the int8R/L midline interface by PKC-3 plotted over time (teal, mean and standard deviation of n=4 embryos where t=0 is clear vir-int contact) and total, bleach corrected PKC-3 intensity across int8 normalized to initial timepoint imaged (grey).

(C) Cyst-to-tube transition (PKC-3, magenta, gamma=0.5 for simultaneous apical and nuclei visualization) in intestine (pha-4p::PM, green) after laser ablation of indicated progenitor (nhr-67p::H1, magenta, gamma=0.5).

(D) Normalized int8 PKC-3 intensity plotted as mean (p-value, t-test).

(E,F) Degradation of ZF degron-tagged E-cadherin (HMR-1, green) in nhr-67+ cells (blue (dotted outline)) shown in maximum intensity projection. Magenta bracket, apical gaps in PKC-3 (magenta). Arrow, abnormal rectum.

(F) Intestinal apical gaps binned by degradation strength as in Figure S4I. p value=0.004, Fisher’s Exact Test. G) Cyst-to-tube transition time courses. Apical surface did not extend in n=0/8 control and n=2/6 HMR-1^nhr-67p(−) embryos. Controls in E-G lack degradation transgene.

All images unless noted are lateral 3D renderings. Scale bars, 5 μm.

See also Figure S4 and Video S3.

Next, we tested the role of cell-cell contact between the vir cells and the intestine in guiding the localization of apical proteins during the cyst-to-tube transition. The homophilic transmembrane adhesion protein E-cadherin underlies cell-cell contact in many types of epithelia^44,45. In C. elegans, the sole E-cadherin homolog HMR-1 has been implicated in tissue connections: HMR-1 is localized to cell-cell contacts during pharyngeal valve and spermatheca-uterine valve morphogenesis^46,47, required to align apical surfaces along the polarizing intestine⁴⁸, and zygotic hmr-1 mutants occasionally show a disconnection between the pharynx and the epidermis⁴⁹. We therefore tested whether E-cadherin in the vir cells is required during vir cell contact with the intestine to drive the intestinal cyst-to-tube transition. To degrade HMR-1 in vir cells prior to their contact with the intestine, we used a tissue-specific degradation system in which the target is endogenously tagged with a ZF degron and the degrader protein ZIF-1 is expressed in a cell or tissue of interest^50,51. We used a ZF-tagged allele of HMR-1^48,52 and expressed ZIF-1 with the promoter for nhr-67 (nhr-67p, HMR-1^nhr-67p(−)), a gene which is expressed in the vir precursor cells as well as other ABp-derived cells^53,54 (Figure S4F–G). Although this approach yielded incomplete HMR-1 depletion, a small but consistent minority of 1.5-fold stage HMR-1^nhr-67p(−) embryos showed intestinal defects (Figure 3E–F, Figure S4H–I). While the intestinal primordium was generally polarized in HMR-1^nhr-67p(−) embryos, we observed breaks in the apical surface at the posterior end of the intestine, suggesting a potential failure in the cyst-to-tube transition. Live imaging of this process in HMR-1^nhr-67p(−) embryos revealed an inability of apical proteins to spread ventrally between the left-right int8 surfaces (Figure 3G). Even given the incomplete HMR-1 depletion, these observations are consistent with a role for E-cadherin within vir cells to signal the reorientation of the apical surface in intestinal cells that underlies the cyst-to-tube transition. Together, these data show that vir cell contact is required for the first step in joining the intestinal lumen to that of the rectum: extending and exposing the apical surface at the posterior end of the intestine.

Rectal valve cells are bipolar

Once the apical surface of the intestine is exposed and therefore competent to join with the apical surface of the neighboring tissues, the next morphogenetic step is to align and connect the rectal and intestinal apical surfaces (Figure 2C–F). We therefore explored how vir cells join the end of the intestinal apical surface to the tip of the rectal lumen. Live imaging showed puncta of apical and junctional proteins at the intestine-valve and valve-rectum interfaces that merged during embryonic elongation from the 1.5- to 2-fold stage, filling in the gap in apical continuity (Figure 2D, Figure 2F, Figure S3). Due to the resolution limits of our imaging, we were unable to determine the extent to which these puncta were localized within the vir cells as opposed to within the neighboring intestinal and/or rectal tissues (Figure 1A). We therefore investigated if vir cells themselves contribute to this punctate apical signal by creating PAR-6^vir, an apical reporter inserted into the valve-specific endogenous fos-1a locus^53,55 (Figure S5A–B, Methods). PAR-6^vir localized to an intestine-facing punctum and a rectal-facing punctum (black and white arrowheads, respectively, Figure 4, Figure S5D–E) and overlapped with endogenous PKC-3 signal (Figure 4A’) and weak actomyosin signal (Figure S1B, Figure S5C). Super-resolution imaging of embryos showed two substructures within both the intestine-facing and rectum-facing puncta on thin membrane extensions (Figure 4D, Figure S5F–I), suggesting that both virR and virL had anterior and posterior-facing apical puncta. Substructures were occasionally visible on either side of the virR/L interface within the intestine-facing puncta (Figure 4C–D, Figure S5H). Although membrane signal was often not visible within puncta, the two substructures were parallel to visible virR and virL membranes (Figure S5I). Additionally, EM of 3-fold embryos (R.M. Durbin, Ph. D. thesis, University of Cambridge, 1987) showed that vir cells have thin contralateral contacts with both intestinal cells and that both vir cells contact the rectal lumen (Figure S1F, Figure S5J). These results are most consistent with both vir cells having bipolar and not antipolar localization of apical proteins prior to the completion of apical connectivity (Figure 4C,E).

Figure 4: Rectal valve cells are bipolar.

(A) PAR-6^vir (black) shows apical surface of 3 FOS-1a+ embryonic cells (one pharyngeal valve (arrow) and vir cells (arrowheads)), colocalizing with rectum-facing (white arrowhead) and intestine-facing (black arrowhead) PKC-3 (magenta) puncta. (A’) Line scans (double line in cartoon) of PAR-6^vir (green) and PKC-3 (magenta) background-subtracted intensity at both vir puncta (mean and 95%CI of n=12 1.5-fold embryos) (A’’) Time course of embryo in A aligned to rectum. Overlap of PAR-6^vir (green) and PKC-3 (magenta) signal yields black. Single z-slices upon twitching at 160 minutes.

(B, D) PAR-6^vir (black) with cells recolored by tissue^† (end-1p::PM/nhr-67p::PM timecourse, his-72::PM super-resolution).

(C) Potential models for PAR-6^vir.

(D) Optical 280X super-resolution images of PAR-6^vir substructures (brackets; 2 substructures visible in 14/14 rectum-facing and 16/16 intestine-facing puncta). Insets show 0.3μm z-slices from locations indicated in 3D models of membrane signal shown with separation between the two cells in D’. Scale bars, 1 μm.

(E) Example 100x images of PAR-6^vir in CO₂-paralyzed embryos with model of vir puncta over time from n=215 embryos.

All images, unless noted, are maximum intensity projections of 1.5-fold embryos. Scale bars, 5 μm.

See also Figure S5.

Following the appearance of the bipolar distribution of polarity proteins, the intestinal and rectal-facing puncta merged into a single apical surface along the virR and virL interface (Figure 4A,E, Figure S5D–E). Near embryonic hatching (late 3-fold) and in the L1 larval stage, the PAR-6^vir signal became weaker and more unpolarized. Endogenously-tagged PAR proteins were not routinely visible in vir cells in L1 larvae, and junctional proteins localized in a pattern consistent with a dorsal and ventral junction between the two vir cells (Figure S5D,K–N). Therefore, the joining of the internal intestinal apical surface and the external-facing rectal lumen occurs by bipolar cells merging apical surfaces along the virR/L interface during embryogenesis.

Intestinal cells are required for valve bipolarity

Given the bipolar localization of polarity puncta within the vir cells, we hypothesized that these cells gain polarity information from the tissues that they connect. We therefore asked whether the intestine is required for vir cell polarization. Using laser ablation, we killed the precursor to the posterior intestinal cells before the vir cells contacted the posterior intestine (Figure 5A). As expected, the ablated intestinal cells remained as corpses within the embryo but lacked any apical localization of polarity proteins. Strikingly, although cells appeared to contact the intestinal cell corpses, no intestine-facing PKC-3 positive puncta were visible (Figure 5A), indicating that live intestinal cells are required to induce vir polarity at this interface. To further test whether intestinal cells are required for vir polarity, we leveraged a severe phenotype resulting from intestine-specific degradation of the PP2A phosphatase scaffolding subunit⁵⁶ PAA-1 (PAA-1^gut(−)) in which intestinal cells perdured in mitosis for hours instead of minutes and hatched with an intestinal cyst disengaged from surrounding tissues (Figure S6A–B). PAA-1^gut(−) embryos were extremely delayed in intestinal polarization (comma – 1.7-fold embryos with disconnected apical patches: control: n=0/18, PAA-1^gut(−): n=18/18), allowing us to ask whether alive but severely compromised intestinal cells induce vir bipolarity. In PAA-1^gut(−) embryos, vir cells appeared to contact and push into the posterior intestine, yet no intestine-facing puncta were visible and instead all apical signal redistributed to the rectal-facing puncta (Figure 5B–C). PAA-1^gut(−) intestinal cysts represent a severe phenotype that is a sum of prolonged mitosis, cell shape changes, reduced apical polarization and adhesion, among many other potential effects; therefore, we tested roles for mitosis (Figure S6C–L) and for individual apical and cell-cell adhesion molecules (Figure 6, see below) in vir cell polarization.

Figure 5: Valve cell polarity is informed by signaling from the intestine.

(A) Apical signal after posterior intestinal progenitor laser ablation (black arrowhead; vir-intestine contact determined from global pie-1p::PM; no ablation: n=10, Eprp/Eplp ablation: n=13).

(B) Apical surfaces (PKC-3, black) in embryos with intestine-specific depletion of PP2A phosphatase subunit PAA-1 (blue).

(C) Percentage of total PKC-3 intensity in vir partitioned between apical puncta. Right, total intensity as percentage of mean of control embryos (dashed line).

(D) Intestinal mitotic perdurance (end-1p::PM, elt-2p>N (bipartite transgene labeling nuclei (N), see Methods), green) in control (no degron-tagged allele) and PLK-1^{E8 gut(−)} embryos. *, uncertainty in apical (PKC-3, black) intestine-facing puncta location.

(E) Model for valve polarity after mitotic perdurance.

(F) Tissue-specific apical signal in PLK-1^{E8 gut(−)} embryos. Black arrow, PAR-6^int and PAR-6^vir signal present at posterior end of intestinal cells.

(G) PAR-6^vir signal quantified as in (C). Note that rectum-facing percentage equals 100% - intestine-facing percentage.

Unless noted, all images are maximum intensity projections of 60X confocal images of 1.5-fold embryos, quantifications from comma—2-fold embryos, and controls lack degradation transgene. Mean +/− standard deviation. p-values, Mann-WhitneyU. Scale bars, 5 μm.

See also Figure S6.

Figure 6: Intestinal E-cadherin is required for valve polarity.

(A-B) ZF_degron::GFP-tagged protein localization at the intestine-facing (black arrowhead) and rectum-facing (white arrowhead) puncta in control and indicated gut(−) embryos.

(C-D) Apical (PKC-3) and E-cadherin/HMR-1 signal in HMR-1^gut(−) embryos. Circled point in HMR-1^gut(−) graph indicates embryo without intestinal apical gaps.

(E-F) PAR-3 and HMR-1 intensity with control (gut(−) with untagged maternal) that isolates signal from the intestine-facing puncta.

(G,H) HMR-1 localization and quantification in vir after β-catenin/HMP-2 depletion. (G’) scaled to show intestine-facing signal. In H, datapoints for each embryo in graph on left are colored by ratio of junctional (AJ) versus lateral (lat.) HMR-1 intensity in the intestine.

All images are maximum intensity projections of 1.5-fold embryos, controls lack degradation transgene, and quantifications as in Figure 5C. Mean +/− standard deviation. p-values, Mann-Whitney[IPS]U. Scale bars, 5 μm.

See also Figure S7.

Mitosis challenges apical continuity between neighboring cells, including in the C. elegans intestine^57–61. We asked whether prolonged intestinal mitosis impacts vir bipolarity by inducing mitotic perdurance in intestinal cells (Figure 5D–E, Figure S6C–L) and assaying vir- (PAR-6^vir) and intestine-specific (PAR-6^int) apical surfaces separately (Figure 5E–G, Figure S6M–N). We disrupted mitosis in the intestine by depleting mitotic regulators starting at E8 (‘E8 gut(−)’, ifb-2p::ZIF-1⁵¹) prior to the E8-E16 division that precedes intestinal polarization^4,38. Depletion of kinases CDK-1, AIR-1/Aurora A, PLK-1/Polo-like, centrosomal protein SPD-5, or PAA-1 resulted in fewer intestinal cells, perdurance of cell rounding and mitotic figures, and/or apical gaps that worsened during embryo elongation particularly in the posterior (Figure S6E–K). In PLK-1^{E8 gut(−)} embryos, intestinal apical polarization was prolonged, potentially impacting the initial cell-cell contact between the posterior intestine and vir cells (Figure S6L); however PAR-6^vir signal was almost always weakly present at the end of the intestine (Figure 5F, Figure S6M–N). Both PAR-6^int and PAR-6^vir signal were maintained even when the intestinal apical surface became discontinuous during elongation (2-fold, Figure S6M–N) suggesting that mitotic perdurance alone was not sufficient to disrupt apical continuity between these neighboring cells once it was established. However, the intestine-facing PAR-6^vir signal was diminished and redistributed to the rectal-facing puncta, indicating that intestinal mitotic perdurance attenuated vir polarization (Figure 5F–G). This redistribution of PAR-6^vir between interfaces likely reflects repartitioning between apical sites within the same cell, further supporting the model that vir cells are bipolar (Figure 5E). We hypothesize that the attenuation of vir polarity at the intestine-facing puncta might be due to the reduction and/or remodeling of apical and junctional proteins that normally occurs during mitosis⁶¹, thus we next tested a role for these proteins in the intestine in establishing vir cell polarity.

Intestine-specific E-cadherin is required for valve cell bipolarity

To identify the molecular mechanisms by which intestinal cells mediate the formation of bipolarity in the neighboring vir cells, we performed a targeted screen of apical and adhesion proteins. Using tissue-specific degradation, we depleted proteins from the embryonic intestine prior to intestinal polarization (gut(−), elt-2p::ZIF-1^27,48,61) and then examined the vir-intestine interface. Given that our degradable alleles were endogenously tagged with GFP, this screen allowed us to visualize the endogenous localization at the intestine-facing puncta in the vir cells (Figure 6A, black arrowhead), confirming PAR-6^vir localization. Indeed, all apical and junctional proteins tested localized to intestine-facing puncta within the vir cells that were highlighted against the loss of intestinal signal (Figure 6A–B).

Removal of apical PAR complex proteins from the intestine had no effect on the corresponding intestinal-facing puncta of these proteins in the vir cells, despite severely compromising the intestinal apical surface^27,48,61. Similarly, degradation of junctional components with previously validated alleles^27,48 had little effect (Figure 6A–B), although DLG-1 localized inconsistently to the puncta even in wild-type embryos. The striking exception was the depletion of E-cadherin/HMR-1. HMR-1^gut(−) embryos had a strong reduction in both HMR-1 and PKC-3 at the vir-intestine interface, indicating a significant reduction or loss of the vir puncta (Figure 6C–D). To quantify the vir portion of endogenous HMR-1 in the vir puncta, we compared HMR-1^gut(−) embryos with those rescued by maternal (non-fluorescent, non-degradable) HMR-1. This control allowed us to compare endogenous HMR-1 levels in vir at the intestine-facing puncta with and without functional intestinal HMR-1 (Figure 6E–F). The sum of the HMR-1 fluorescence intensity of the bipolar rectal-facing and intestinal-facing vir puncta remained constant with or without functional intestinal HMR-1, consistent with the intestine-facing and rectal-facing puncta drawing from a shared pool of protein components in vir to correctly partition bipolarity as seen following other intestinal perturbations (Figure 5). Conversely, degradation of the key polarity determinant PAR-3 did not affect vir bipolarity (Figure 6E–F). We previously found that HMR-1 is required to polarize the intestine, independent of its conserved junctional partners α- and β- catenin, pointing to a signaling rather than junctional role⁴⁸. Similarly, α-catenin/HMP-1^gut(−) or β- catenin/HMP-2^gut(−) had normal valve polarity (Fig. 6A–B). To rule out a junctional contribution for intestinal HMR-1 in vir polarization, we first verified that HMP-2^gut(−) embryos had the reported disorganization of intestinal apical junctions at comma/1.5-fold stage⁴⁸ (Figure S7). Despite a strong effect on intestinal HMR-1 junctional localization, HMP-2^gut(−) embryos showed no change in the intensity of the vir HMR-1 intestine-facing puncta, and [HMR-1;HMP-2]^gut(−) embryos did not further reduce vir HMR-1 intensity compared to HMR-1^gut(−) embryos alone (Figure 6G–H). These results point to E-cadherin-mediated signaling at the cell-cell contact as the crucial factor coordinating apical alignment across different tissues, implying that E-cadherin, independent of catenins, can inform apical localization in neighboring cells.

Discussion

Here we show that cells at the boundaries of connecting epithelial tissues reciprocally exchange orientation information to correctly align their apical surfaces. Using the morphogenesis of the posterior C. elegans digestive tract, we defined how a specialized two-cell tissue, the rectal valve, enables the internal intestinal lumen to join the external-facing rectal lumen via E-cadherin-mediated cell-cell interactions (Figure 7). The process requires two steps: the endodermal intestine transitions from a cyst to a tube and then the exposed intestinal apical surface joins to the ectodermal rectum by merging of bipolar apical surfaces within vir cells.

Figure 7: Reciprocal signaling between tissues connects the digestive tract.

Summary of normal intestinal-rectal connection via reciprocal E-cadherin-mediated signaling between vir and posterior intestinal cells.

The reorientation of existing polarity within the posterior intestine involves the extension of apical surfaces along the cell-cell interface to connect with vir cells. This reorientation of existing apical and junctional surfaces where endodermal tissue abuts ectodermal tissue, such as those observed in pharyngeal arches⁶² and foregut⁶³, is a widespread feature of development that is not well understood. Our data point to separable pathways for polarity establishment and tissue connections, as PAR-3^gut(−) embryos displayed normal vir bipolarity, despite drastic apical polarity defects within the intestine^25,27,38,48. Connections at tissue boundaries in systems from vertebrate kidneys to Drosophila testes similarly appear to be under separate genetic control than the initial process of apical polarization^49,64–68. However, future studies using tissue-specific approaches will be required to identify general principles of how polarity establishment interfaces with pathways controlling tissue connection.

Bipolarity within cells at tissue interfaces may be a common way to mediate tissue connections. Although we cannot conclusively rule-out that vir cells are antipolar, our data strongly suggest that vir polarization begins from a bipolar distribution of apical determinants. The bipolar puncta then draw together to join intestinal and rectal surfaces to create a continuous lumen. Where the force that draws these surfaces together comes from is currently unknown, however a model where actomyosin contractility between the puncta facilitates merging⁶⁹ is inconsistent with our localization data. Future work will be necessary to test whether the puncta merging forces arise from neighboring apical domains pushing⁷⁰, vertex sliding⁷¹, and/or vesicle-based lumen extension^72,73.

Vir cells establish bipolarity from an unpolarized state as they integrate with the already polarized intestine, transitioning along the mesenchymal-epithelial continuum. This integration is similar to radial intercalation of unpolarized cells into existing epithelial tubes and sheets^74–77 except that vir cells experience competing sources of apical orientation information from contact with both intestinal and rectal cells. During lumen formation, polarizing cells at the edges between cysts similarly find themselves in a bipolar environment before apical alignment is complete^78–82. Cells at tissue-tissue boundaries often integrate competing sources of polarity information, whether they lose polarity and then repolarize as in kidney and optic fissure closure^39,83 (analogous to early vir polarization upon intestinal contact) or they maintain aspects of polarity while simultaneously invading into the neighboring tissue as in avian lung, mouse axial mesoderm, zebrafish intestine, regenerating zebrafish kidney, and ascidian branchial fissures^{62,64,68,74,84} (analogous to later bipolar vir cells). Vir cells polarize with input from the intestine via E-cadherin signaling, likely via cell-cell contact independent of its role in catenin-mediated adhesion⁴⁸. Thus, vir cells are a model for the formation of this understudied bipolar state during tissue connection.

We speculate that developmentally appropriate bipolarity, putatively reflecting a partial mesenchymal-partial epithelial state, may be hijacked by carcinomas to drive metastatic cell behaviors, as cell clusters and cells with partial epithelial-mesenchymal transitions are the most likely drivers of productive metastases^85–87. Additionally, E-cadherin is present on invadopodia^88,89, suggesting that signaling at cell-cell contacts and the transfer of apical orientation information is important in not only developmental but also disease contexts.

Resource Availability

Lead contact

All requests for information and resources should be directed to the lead contact, Jessica L. Feldman (feldmanj@stanford.edu).

Materials availability

All requests for plasmids and strains generated in this study should be directed to and will be fulfilled by the lead contact.

STAR Methods

Experimental model and study participant details

Husbandry of experimental animals

C. elegans strains were cultured on Nematode Growth Medium (NGM) plates coated with a lawn of E. coli OP50⁹⁰ unless otherwise stated. For all experiments, animals were maintained at 20°C. Strains used in this study are listed in the Key Resources Table and listed by figure in Methods S1. For all experiments, embryos were obtained from hermaphrodites. Embryos were either harvested from the uterus of young day 1-2, gravid adult hermaphrodites that were incubated in M9 media or were picked directly from culture plates after they were laid by hermaphrodites. For strain generation that required genotypic mixing, male and hermaphrodite strains were crossed. 5 adult male worms and 3 L4 hermaphrodite were incubated on 3.5 cm NGM plates seeded with OP50 and cross-progeny were isolated. To create homozygous strains, single F2 progeny were isolated and confirmed to be homozygous or alternatively, cross-progeny males were mated to GRP-1 overexpression strains⁹¹ PD2218 or PD2217 and single non-Mendelian progeny were isolated. In all cases, the desired genotypes were identified either through direct visualization of markers when possible or by PCR genotyping.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
anti-Phospho Histone H3 (Ser10) Antibody	Cell Signaling	9701S
Donkey Anti-Rabbit-Alexa Fluor 647 - AffiniPureIgG (H+L)	Jackson ImmunoResearch Laboratories	711-605-152
Bacterial and virus strains
E. coli OP50	CGC	OP50
Chemicals, peptides, and recombinant proteins
Levamisole hydrochloride	Sigma	Cat#31742
Alt-R S.p. Cas9 Nuclease V3	Integrated DNA Technologies	1081058
FD and C Blue #1 Powder	FLAVORSandCOLOR.com	FD&C Blue #1
Critical commercial assays
Q5 Site-Directed Mutagenesis Kit	New England Biolabs	Cat#E05548
NEBuilder HiFi DNA Assembly	New England Biolabs	Cat#E5520S
Deposited data
Segmentations from Guan et al 2025	Guan et al37	https://bcc.ee.cityu.edu.hk/cmos ; https://doi.org/10.6084/m9.figshare.24768921
imaging chamber	NIH 3D	3DPX-022973
16 hr L1 larvae slices 847, 868, 887 from Mulcahy et al 2022	Mulcahy et al29	https://doi.org/10.60533/BOSS-2022-1JI1 ; https://bossdb.org/project/mulcahy2022 ; https://wormfindr.web.illinois.edu/v1/
bean (320mpfc) EM dataset from Santella et al. 2022	Santella et al108	webKnossos ID 1712,526,428,0,2. 2,576. C. elegans bean stage EM. https://wklink.org/1322
comma (345mpfc) EM dataset from Santella et al. 2022	Santella et al108	webKnossos ID 399,796,779,0,4.7 16,424. C. elegans comma stage EM. https://wklink.org/1623
~550 min 3-fold EM dataset from Durbin 1987	wormatlas.org	RDE
this paper: processed images for figures	Mendeley Data repository	10.17632/7n37yyb9vz.4
this paper: raw z-stacks from all figure images	Mendeley Data repository	10.17632/7n37yyb9vz.4
this paper: code for plots and source data	Mendeley Data repository; gitbhub	10.17632/7n37yyb9vz.4 ; https://github.com/feldworms/vir_paper_Cote2026
this paper: plasmid and allele maps	Mendeley Data repository	10.17632/7n37yyb9vz.4
Experimental models: Organisms/strains
pkc-3(wow118[RFP::PKC-3])27 II; cs262[let-653::SfGFP]109 IV; pxIs10[pha-4::GFP::CAAX, rol-6]63	this study	JLF1273
pkc-3(it309[GFP:pkc-3])110 II ;zif-1(gk117)51 III; Itis37[pie-1p::mCherry::his-58; unc-119(+)]111 IV; ItIs44[pie-1p::mCherry::PH(PLC1delta1); unc-119(+)]112 V; eps-8(wow143[eps8A:mScarlet:Flag])27 IV; stIs10440 [ceh-27::H1-wCherry::his-24::mCherry + unc-119(+)]54; stIs10116[his-72p::his-24::mCherry::let-858 3’ UTR; unc-119(+)]	this study	JLF1107
jsTi1493 wowSi10[nhr-67::mScarlet::CAAX] IV; fos-1(bmd138[gfp:fos-1])95 V; lam-2(qy20[LAM-2::mN])113 X; wowEx237 [end-1::PH::mCherry, rol6]	this study	JLF1402
pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III; pxIs10[pha-4::GFP::CAAX + rol-6(su1006)]; stIs10684 [nhr-67::H1-wCherry + unc-119(+)] 54	this study	JLF914
zif-1(gk117) III; jsTi1493 wowSi31[ceh-27::LifeAct::mScarlet] IV; zuIs70[end-1::GFP::CAAX]35 V	this study	JLF1389
pkc-3(wow118[RFP::PKC-3]) II ;let-653(cs262[let-653::SfGFP]) IV	this study	JLF1272
pxIs10[pha-4::GFP::CAAX]; stIs10684[nhr-67::H1-wCherry + unc-119(+)]; TH110[pie-1::PAR-6] 114	this study	JLF684
zif-1(gk117) III; jsTi1493 wowSi10 [nhr-67:mScarlet-I:CAAX] IV; pxIs10 [pha-4:GFP:CAAX, rol-6]	this study	JLF1267
pkc-3(wow118[RFP::PKC-3]) II; ncam-1(wow107(NCAM-1::ZF::GFP))48 X	Naturale et al48 (equivalent to JLF756)	JLF1674
pkc-3(wow118[RFP::PKC-3]) II; fos-1(wow177[ZF::BFP::FOS-1])/nT1 V; lam-2(qy20[LAM-2::mNeonGreen]); wowEx128[nhr-67::ZIF-1::E2A::BFP::CAAX::unc54 + myo2::mCherry]	this study	JLF1469
jsTi1493 wowSi31[ceh-27p::Life::mScarlet] IV; lam-2(qy20[LAM-2:mNG]) X	this study	JLF1391
zif-1(gk117) III ; ncam-1(wow107[NCAM-1:ZF:GFP]), lam-2(qy41[LAM-2:mKate])113 X	this study	JLF1378
zif-1(gk117) III; let-413(wy1413[let-413::aid::gfp])52 V; dlg-1(wow141[dlg-1:mScarlet])27 X	Pickett et al27	JLF1176
[let-413::gfp+rol6(su10060)]21 par-3(it300[par-3::mCherry]) III	Pickett et al27	JLF309
hmr-1(wy1254[HMR-1:ZF:GFP])48 I; pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III	this study	JLF752
pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III; pxIs10[pha-4::GFP::CAAX + rol-6(su1006)]	this study	JLF913
hmr-1(wy1254[HMR-1:ZF:GFP])/hT2 I; pkc-3(wow118[RFP::PKC-3]) II; ; zif-1(gk117)/hT2 III; wowEx128[nhr-67p::vhhGFP::ZIF-1::E2A::BFP::CAAX]	this study	JLF1817
hmr-1(wow193[hmr-1::mScarlet-I])48 I; par-3(it298[par-3::gfp]) III; wowEx52[end-1p::bfp::caax; rol-6(su1006)]	Naturale et al48	JLF1506
spd-5(wow53[zf:gfp:spd-5]) 99 /hT2 I; zif-1(gk117)/hT2 III; pkc-3(wow118[RFP::PKC-3]) II; wowEx128[nhr-67p::vhhGFP::ZIF-1::E2A::BFP::CAAX]	this study	JLF903
pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III; fos-1(wow181[PAR-6:mNeonGreen:P2A:fos-1]) V ; wowEx128[nhr-67p::vhhGFP::ZIF-1::E2A::BFP::CAAX]	this study	JLF1440
zif-1(gk117)III; jsTi1493 wowSi10 IV; fos-1(wow181[PAR-6:mN:P2A:fos-1]) V; wowEx237 [end-1::PH::mCherry, rol6]	this study	JLF1426
zzyIs139 [his-72p::PH(PLC1delta1)::mCherry::pie-1 3’ UTR + unc-119(+)]37 II; fos-1(wow181[PAR-6:mN:P2A:fos-1]) V	this study	JLF1865
cap-1(cp436[mScar::CAP-1])115 IV; fos-1(wow181[PAR-6:mN:P2A:fos-1]) V	this study	JLF1798
zif-1(gk117) III; fos-1(wow181[PAR-6:mN:P2A:fos-1]) IV; dlg-1(wow185[DLG-1:mScarlet-I-Glo]) X	this study	JLF1424
zif-1(gk117); jsTi1493 wowSi40[elt-2::PAR-6::tagRFP] IV; fos-1(wow181[PAR-6:mN:P2A:fos-1]) V	this study	JLF1776
pkc-3(it309[gfp::pkc-3]) II; Itis37[pie-1p::mCherry::his-58; unc-119(+)] IV; ItIs44[pie-1p::mCherry::PH(PLC1delta1); unc-119(+)] V; stIs10116[his-72p::his-24::mCherry::let-858 3’ UTR; unc-119(+)]	Naturale et al48	JLF640
pkc-3(wow118[RFP::PKC-3]) II; paa-1(wow140[ZF:GFP:PAA-1]), zif-1(gk117)/hT2 III; wowEx111[elt2p::zif-1, end-1::BFP::CAAX, myo2p::mCherry] 116	this study	JLF1034
pkc-3(wow118[RFP::PKC-3]), jsSi1579 wowSi41[elt-2p>GFPnls] II; zif-1(gk117) III; wowIs3[ifb-2::zif-1, myo-2::mCherry] 51 , zuIs70[end-1::GFP::CAAX] V	this study	JLF1527
pkc-3(wow118[RFP::PKC-3]), jsSi1579 wowSi41[elt-2p>GFPnls] II; plk-1(wow169[PLK-1::ZF::GFP11]), zif-1(gk117)/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1560
plk-1(wow169[PLK-1::ZF::GFP11]), zif-1(gk117)/hT2 III; jsTi1493 wowSi40[elt-2::PAR-6::tagRFP] IV; wowIs3[ifb-2::zif-1, myo-2::mCherry], fos-1(wow181[PAR-6:mN:P2A:fos-1]) V	this study	JLF1778
plk-1(wow169[PLK-1::ZF::GFP11]), zif-1(gk117) III; jsTi1493 wowSi40[elt-2::PAR-6::tagRFP] IV; fos-1(wow181[PAR-6:mN:P2A:fos-1]) V	this study	JLF1781
zif-1(gk117) III; jsTi1493 wowSi40[elt-2::PAR-6::tagRFP] IV; fos-1(wow181[PAR-6:mN:P2A:fos-1]) V	this study	JLF1776
unc-119(ed3) III; wgIs304[fos-1::TY1::EGFP::3xFLAG(92C12) + unc-119(+)] 117	Sarov et al117, CGC	OP304
wyEx8629[Pdat-1::HisCl; Podr-1::GFP]; fos-1(wy1539[fos-1::GFP]) 118	Yee et al118	TV27774
fos-1(bmd138[gfp:fos-1a])	Medwig-Kinney et al95, CGC	DQM497
pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III; nmy-1(mc82[nmy-1::GFP])119 X	this study	JLF1931
nmy-2(cp69[nmy-2::mkate2 + LoxP]) I; zif-1(gk117) III; fos-1(wow181[PAR-6:mN:P2A:fos-1]) V	this study	JLF1932
eps-8(wow143[eps8A:mScarlet:Flag]) IV; cutl-18(syb8442[mNG::cutl-18])120,121 V	this study	JLF1863
par-6(wow31[PAR-6::ZF::GFP])61 I; zif-1(gk117) III	this study	JLF194
paa-1(wow139[ZF:GFP:PAA-1])/ht2 I; ItSi910(intDeg)122 II; zif-1(gk117)/hT2 III	this study	JLF925
hT2/+ I; pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117), wow188[ZF::GFP11::CDK-1]/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1474
hT2/+ I; pkc-3(wow118[RFP::PKC-3]) II; plk-1(wow169[PLK-1::ZF::GFP11]), zif-1(gk117)/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1478
pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III; air-1(wow14[AIR-1::ZF::GFP]) 51 , wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1481
pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1480
pkc-3(wow118[RFP::PKC-3]), jsSi1579 wowSi41[elt-2p>GFPnls] II; zif-1(gk117) III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V air-1(wow14[AIR-1::ZF::GFP])/eT1 V	this study	JLF1564
hT2/+ I; pkc-3(wow118[RFP::PKC-3]), jsSi1579 wowSi41[elt-2p>GFPnls] II; zif-1(gk117), wow188[ZFGFP11::CDK-1]/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1561
spd-5(wow192[ZFBFP::SPD-5])/hT2 I; pkc-3(wow118[RFP::PKC-3]), jsSi1579 wowSi41[elt-2p>GFPnls] II; zif-1(gk117) III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1563
hT2/+ I; pkc-3(wow118[RFP::PKC-3]), jsSi1579 wowSi41[elt-2p>GFPnls] II; zif-1(gk117), paa-1(wow139[ZF::GFP::PAA-1])/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry] V	this study	JLF1576
hT2/+ I; pkc-3(wow118[RFP::PKC-3]), jsSi1579 wowSi41[elt-2p>GFPnls] II; zif-1(gk117), paa-1(wow139[ZF::GFP::PAA-1])/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1562
ltSi569[TBG-1::mCherry] I; plk-1(wow169[PLK-1::ZF::GFP11]), zif-1(gk117) III; zuIs70[end-1::GFP::CAAX] V	this study	JLF1479
ltSi569[TBG-1::mCherry]/hT2 I; plk-1(wow169[PLK-1::ZF::GFP11]), zif-1(gk117)/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1292
ltSi569[TBG-1::mCherry] I; zif-1(gk117)/hT2 III; wowIs3[ifb-2::zif-1, myo-2::mCherry], zuIs70[end-1::GFP::CAAX] V	this study	JLF1200
zif-1(gk117), par-3(wow59[PAR-3::ZF::GFP]) 27 /qC1 III; sma-1(wow122[SMA-1:mScarlet-FLAG]) V; wowEx111[elt2p::zif-1, end-1::BFP::CAAX, myo2p::mCherry]	this study	JLF1169
par-6(wow31[PAR-6:ZF:GFP])/hT2 I; zif-1(gk117)/hT2[GFP] III; wowEx10[elt-2::ZIF-1, end-1::mCherry::tba-1, rol-6] 51	this study	JLF356
pkc-3(wow85[ZF::GFP::PKC-3])27 wowIs28[elt-2::ZIF-1, end-1::His::mCh, myo-2::mCherry]116/mIn1 II; zif-1(gk117) III	Pickett et al27	JLF646
pkc-3(wow85[ZF::GFP::PKC-3])61 II; zif-1(gk117) III	Sallee et al61	JLF480
dlg-1(wow117[DLG-1::ZF::GFP]) 61 /szT1 X; afd-1(wow133[AFD-1::mScarlet]) I; zif-1(gk117) III; wowEx111[elt2p::zif-1, end-1::BFP::CAAX, myo2p::mCherry]	this study	JLF828
dlg-1(wow117[DLG-1::ZF::GFP])/szT1 X; zif-1(gk117) III; ltSi910 (intDeg) II	Pickett et al27	JLF829
par-1(wow87[PAR-1::ZF::GFP]) 27 /nT1 I; pkc-3(wow118[RFP::PKC-3]) II; zif-1(gk117) III; wowEx111[elt2p::zif-1, end-1::BFP::CAAX, myo2p::mCherry]	Pickett et al27	JLF827
hmp-1(wow138[HMP-1:ZF:GFP]) 48 /tmC3; zif-1(gk117); wowEx113[elt-2::ZIF-1, end-1::His::mCh, myo-2::mCherry]	Naturale et al48	JLF1015
hmp-2(wow108[HMP-2:ZF:GFP]) 48 /tmC27; zif-1(gk117); wowEx113[elt-2::ZIF-1, end-1::His::mCh, myo-2::mCherry]	Naturale et al48	JLF1279
afd-1(wow106[AFD-1::ZF::GFP]) 27 /hT2 I; zif-1(gk117)/hT2 III; pkc-3(wow118[RFP::PKC-3]) II; wowEx111[elt2p::zif-1, end-1::BFP::CAAX, myo2p::mCherry]	Pickett et al27	JLF862
hmr-1(wy1254[HMR-1::ZF::GFP])/ hT2 I; zif-1(gk117)/hT2 III; pkc-3(wow118[RFP::PKC-3]) II; wowEx111[elt2p::zif-1, end-1::BFP::CAAX, myo2p::mCherry]	Naturale et al48 (equivalent to JLF834)	JLF1089
hmr-1(wy1254[HMR-1::ZF::GFP])/hT2 I; pkc-3(wow118[RFP::PKC-3]), wowIs28[elt-2::ZIF-1, end-1::His::mCh, myo-2::mCherry] II; zif-1(gk117)/hT2 III	this study	JLF1010
hmr-1(cp21[HMR-1::GFP]) 123 , hmp-2(wow238[hmp-2::ZF::mCherry]) I; zif-1(gk117) III	this study	JLF1884
hmr-1(cp21[HMR-1::GFP]), hmp-2(wow238[hmp-2::ZF::mCherry]) I; wowIs28[elt-2::ZIF-1, end-1::His::mCh, myo-w::mCherry] II; zif-1(gk117) III	this study	JLF1879
hmr-1(wy1254[HMR-1::ZF::GFP])/hT2, hmp-2(wow240[hmp-2:mCherry]) I; wowIs28[elt-2::ZIF-1, end-1::His::mCh, myo-w::mCherry] II; zif-1(gk117)/hT2 III	this study	JLF1881
hmr-1(wy1254[HMR-1::ZF::GFP])/hT2, hmp-2(wow239[hmp-2::ZF::mCherry])I; wowIs28[elt-2::ZIF-1, end-1::His::mCh, myo-w::mCherry] II; zif-1(gk117)/hT2 III	this study	JLF1880
ccTi1594 [mex-5p::GFP::gpr-1::smu-1 3’UTR + Cbr-unc-119(+), III: 680195], umnIs7 [myo-2p::GFP + NeoR, III:9421936] III.	Artiles et al91 , CGC	PD2218
ccTi1594mex-5p::GFP::gpr-1::smu-1 3’UTR + Cbr-unc-119(+)] unc-119(ed3 III; hjSi20[myo-2p::mCherry::unc-54 3’UTR] IV	Artiles et al91, CGC	PD2217
See also Methods S1 ‘Strains’ and ‘Alleles and transgenes’
Oligonucleotides
See Methods S1 ‘Oligonucleotides’
Recombinant DNA
See Methods S1 ‘Plasmids’
Software and algorithms
python v3.9.21	Python Software Foundation	https://www.python.org
napari v0.4.19	napari contributors	https://napari.org ; 10.5281/zenodo.3555620
Fiji/ImageJ2 2.16.0/1.54p	Schindelin et al100	https://imagej.net/software/fiji/
scipy v1.13.1	Virtanen et al107	https://scipy.org
tinkercad	Autodesk	https://www.tinkercad.com/

Methods Details

vir marker identification

Databases of expression in lineaged C. elegans embryos^53,54 were mined for expression in ABprpapppp and marker discovery between ABprpapppp and ABplpapppp from lineage-labeled single-cell sequencing data⁵⁵ followed by visual confirmation in embryos yielded nhr-67p (stIs10684), ceh-27p (stIs10440), pha-4p::PM (pxIs10), NCAM-1 (wow107), FOS-1A (bmd138) as markers of vir cells.

Transgenes and endogenous alleles

Cloning strategy for each plasmid, primers, and injection mixes for each transgene or allele detailed in Methods S1.

Endogenously-tagged alleles were created using the self-excising cassette (SEC) CRISPR/Cas9⁹² method except for wow185 DLG-1::mScarlet-I and wow238-240 HMP-2::(ZF)::mCherry that were created via ribonucleoprotein CRISPR/Cas9⁹³. Cloning was performed using Gibson Assembly (NEBuilder HiFi DNA Assembly Master Mix, NEB) using digested backbone and genomic PCR products amplified with touchdown PCR or geneblocks (IDT). The ZF::GFP SEC repair template pJF250⁵¹ was modified to create additional ZF::FP backbones (pLC001 ZF::GFP11, pLC015 ZF::BFP, pLC019 mScarlet::FLAG) and grown in ccdB survival stain (Invitrogen). At least six individual minipreps digested with appropriate enzymes (N terminal tags SpeI/ClaI; C terminal tags AvrII/SpeI) were verified on an 1% agarose gel to not have plasmid recombination before combining for use in HiFi reaction with 0.025 pmol of each component for 4 hours at 45°C to create gene-specific repair templates with 0.5-1kb homology arms. For sgRNA plasmids and pLC150[end-1p::PH::mCherry], site-directed mutagenesis (Q5 SDM kit, NEB) was used to generate plasmids. For SEC-based CRISPR edits, Cas9 and sgRNA were delivered by co-injection with gene-specific plasmid^92,94, isolated by following Rol phenotype before heat-shock to remove SEC, and backcrossed 2x to JLF155 zif-1(gk117) III. To create the PAR-6^vir transgene wow181, the ‘ZF::BFP’ of the SEC repair template for ZF::BFP^SEC^::FLAG::FOS-1a was replaced with ‘PAR-6::mNeonGreen::P2A’ (pLC177) and injected with published sgRNA⁹⁵ so that the fos-1a locus produces separate PAR-6::mNeonGreen and FLAG::FOS-1 proteins after riboskipping.

Plasmids to create single-copy transgenes wowSi15 (elt-2p::tetR-AD) and wowSi31 (ceh-27p::LifeAct::mScarlet) were made with GoldenGate assembly using SapI or BsaI (NEB) ligation^96,97. Single-copy transgenes were made using FLP recombination-mediated cassette exchange⁹⁶. The bipartite transgene labeling intestinal nuclei (‘elt-2p>N’ wowSi41) was made with phiC recombination⁹⁷ of two transgenes at the same chrII locus: elt-2::tetR-AD and tetO::GFP::NLS. Although GFP::NLS would be predicted to be cytoplasmic during mitosis, this allele, which has been PCR verified to have GFP::NLS, is associated with mitotic figures, ie. condensed chromatin-like structures during mitosis potentially due to transient and late nuclear envelope breakdown in C. elegans⁹⁸. Gene diagram of fos-1 locus in Figure S5B from http://www.wormweb.org/exonintron.

Tissue-specific degradation alleles and controls

All degradation strains were maintained on a balancer (with the exception of AIR-1^{E8 gut(−)} and HMP-2^gut(−) with HMR-1::GFP) and all degradation are maternal/zygotic degradation (embryos imaged are from balancer negative mothers). For strains readily viable off of a balancer (all strains except PAA-1^gut(−), PAR complex ^gut(−), and HMR-1^gut(−))^{27,48,51,61,99}, some balancer negative mothers of imaged embryos were 1-2 generations off the balancer. Degradation transgenes to generate gut(−) animals have all been previously validated (see Key Resource Table for references). Degradation in nhr-67p+ cells (neuroblasts, rectal cells) via wowEx128 was verified by using phenotypic or visual readouts: the phenotypic readout was using degradation of SPD-5::ZF::GFP to block mitosis in these cells; visual readout was the quantification of HMR-1 intensity after degradation. Note that transmembrane proteins such as HMR-1::ZF::GFP seem to require higher expression of ZIF-1 to yield strong degradation than other ZF::GFP proteins. Controls for degradation are the same strain lacking the degradation overexpression except for analysis of strains with end-1p::PM and/or elt-2p>N in which ‘control’ is degradation overexpression without the degradable allele. Maternal untagged control in Figure S6A and Figure 6E,F are hT2− embryos from hT2+ mothers who provide untagged maternal protein from the balancer. hT2 is a GFP-marked translocation of chromosome I and III, so GFP-embryos are ¼ euploid embryos with only maternal untagged protein that are able to survive to adulthood and ¾ aneuploid embryos with maternal untagged protein plus zygotic untagged protein from GFP-portion of the translocation with uncharacterized lethality.

Antibody staining

Embryos dissected from one-day old adults of desired genotype incubated in M9 were attached to microscope slides coated with poly-lysine and containing Teflon spacers. Slides were frozen on dry ice, embryos were permeabilized by freeze-crack method⁴ and submerged in 100% MeOH for 5-10 minutes at −20°C. Rehydrated embryos were incubated in 1:500 anti-pH3 (Ser10) overnight at 4°C, washed three times in PBT, incubated in secondary antibody (anti-rb-Alexa-647 1:500) and 500ng/μl Hoescht 33342 for 1 hour at 37°C, washed once in PBT and twice in PBS before mounting in Vectashield (Vector Laboratories).

Embryo and larval imaging

Embryos for microscopy were dissected from gravid hermaphrodites incubated in M9 or collected from unstarved plates containing 1-2-day-old gravid adults. For live imaging, embryo samples were mounted on a pad made of 3% agarose dissolved in M9 under a #1.5 coverslip. Live and fixed imaging was performed on a Nikon Ti-E inverted microscope (Nikon Instruments, Melville, NY) using a 60× Oil Plan Apochromat (NA = 1.4) or 100× Oil Plan Apochromat (NA = 1.45) objective and controlled by NIS Elements software (Nikon). Images were acquired with an Andor Ixon Ultra back thinned EM-CCD camera using 405 nm, 488 nm, or 561 nm imaging lasers and a Yokogawa X1 confocal spinning disk head equipped with a 1.5× magnifying lens. For live imaging, images were taken at a z-sampling rate of 0.5 μm for experiments looking at 1.5-fold embryos and 0.3 μm for experiments targeting bean-stage embryos. For fluorescence imaging of L1 larvae in Figure S5L–N and S6K, L1 larvae were picked into a drop of 2 mM levamisole on a 5% pad to minimize movement. Images were processed the Fiji distribution of ImageJ (“Fiji”) v2.16.0/1.54p¹⁰⁰ or napari v0.4.19.

Laser cell ablations

A Micropoint dye laser (coumarin 435 nm) was targeted at desired cell nucleus (20-50x 50-55% power until debris was raised in DIC) in ventral-mounted embryo. Desired ceh-27p+ or nhr-67p+ cells were identified by location and cell ablation was confirmed by 1) presence of nuclei ejected from the embryo, 2) loss of 2 ceh-27+ cells at 1.5-fold, or 3) loss of nhr-67p++/pha-4p+ wedge-shaped cells at late bean stage. Note that ablation of ABprpapppp before a cell-contact mediated inhibition of vir cell fate in ABplpapppp can result in ABplpapppp transforming into vir cells¹⁰¹. Ablations were performed after ABpxpapppp contact to reduce the impact of cell fate changes on the outcome. To facilitate tracking of ablated embryos, all non-ablated embryos on the pad were exploded with a high power Micropoint laser and embryos were left on the pad overnight at 20°C in a humid chamber. Larvae were imaged as above except without the addition of a paralytic agent before rescuing to a OP50-seeded plate. Rectal length for each larvae was averaged from 3 measurements from single z-slices since larvae were not paralyzed. Larvae one day after laser cell ablation were blind scored as ‘continuous, expanded, or cyst’ and the continuous and expanded categories were combined as ‘connected’ for Figure 1E (data combined from following independent experiments - no ablation: 2 experiments ; ABplpapppp: 1 experiment (examination by EM after PVT progenitor ablation¹⁰² showed that the K.K’ cells moved forward to connect to virL/virR in place of rectD); ABprpapppp: 3 experiments)

Imaging chamber for CO₂ immobilization

The imaging chamber for CO₂ immobilization was designed using AUTODESK Tinkercad (https://www.tinkercad.com/) to improve upon previous embryo CO₂ immobilization setups¹⁰³. The 3D printing files (.stl) are available at NIH 3D 3DPX022973. The chamber comprises two parts: a bottom part with a square chamber that transforms into a bottom coverslip chamber connected to an exit and entry tubing, surrounded by a moat to house an O-ring; and a top part that seals the chamber’s O-ring. The device was printed on a Digital Light Printing 3D printer Photon Ultra (Anycubic, Inc.) using a UV curing stereolithography (SLA) resin. The printing process involved 3 seconds of on and 1 second of off exposure with a z-step of 50 μm and an XY pixel size of 80 μm. After printing, the two parts were cleaned with 100% isopropanol, and a final cure was performed using a UV blue-light LED for 15 minutes. The device was then dried in an oven for 30 minutes. The top and bottom imaging windows were added and sealed using UV clear resin and 22x40mm #1.5 coverslips. A nitrile rubber O-ring (32x28x2 mm) was added to the chamber’s moat to create a seal. Both the top and bottom parts were lined up with N52 5x2 mm round magnets. Embryos were applied to the coverslip in M9 buffer and overlaid with 3% agarose pad. For immobilization, CO₂ was allowed to flow through an ozone water humidifier and into the chamber for 5-10 minutes at a rate of 60-80 cc/min until immobilization. The chamber was then sealed by pinching the tubing at the entry and exit valve. After the imaging was complete, ambient air was inserted into the chamber using an aquarium air pump through a flow derivation system.

SoRa super-resolution imaging

Optical super-resolution imaging was performed on a Nikon CSU-W1 spinning disk confocal microscope with a 100X 1.35NA silicon oil lens and 2.8X SoRa microlens emission disk to yield an image with voxel sizes of 0.03μm x 0.03μm x 0.25μm followed by Richardson-Lucy 3D deconvolution using 10 iterations. Membrane signals were first assigned to cells at a mid-plane and cells were identified by comparison with published Guan et al cell membrane maps³⁷ and then followed through z, especially to identify membranes oblique relative to plane of imaging. Membranes were manually labeled. Line scans were drawn through brightest points of PAR-6^vir signal and then membrane intensity was measured along the same line.

Larval assays

For larval growth assay, 10 young balancer negative mothers of desired genotype were placed on small plates for 4 hrs at 20°C in triplicate to lay eggs on day 0. On day 2/3, larvae were counted and L4s/adults were counted and removed. For Smurf assay, L1 larvae were incubated for 3-5 hours in a humid chamber in 30μl OP50 + 10μl 20% blue dye before mounting and scoring by eye. Example larvae for the ‘Smurf’ feeding assay in Figure S6I were imaged using a Vankey Cellphone Telescope Adapter Mount (Amazon) was used with an Apple iPhone 12mini camera and the compound microscope 60× objective. Brightness, contrast, and hue were adjusted for each color image with Adobe Photoshop v21.0.1.

Phenotypic categorization

Number of 1.5-fold embryos with gap between rectum and intestine was determined from maximum intensity projections. Images were visually inspected for arcade cell connecting at the buccal cavity and clear gap between the intestine and rectum indicating that this apical surface is the final surface to become continuous. Embryos examined from all ‘wild-type’ and control conditions containing the following endogenously-tagged alleles of PKC-3 (wow118, it309, wow85, cp41), PAR-3 (wow59, it300), HMR-1 (wy1254, wow193), or DLG-1 (wow185, wow117).

For mitotic regulator characterization: ‘Apical gaps’ in embryos were scored visually by looking for absence of signal along the midline, as determined from plasma membrane signal, of the anterior-most or posterior-most cells. The plasma membrane marker zuIs70 is often associated with 21-cell intestines (19% of control L1) due to an extra division in the mid-intestine after E20^28,104,105. ‘Mitosis’ was scored when at least one of the L/R pair of cells appeared rounded and/or had condensed mitotic figures. ‘1-cell’ or ‘2-cell’ intestinal ring was scored using plasma membrane marker surrounding the most posterior intestinal nuclei. For Figure S6N and 6B: intestine-facing puncta presence/absence was scored from maximum intensity projections through the intestine with auto-adjusted brightness/contrast (saturated=0.35).

Localization patterns

Localization patterns for endogenously-tagged proteins in non-disrupted genetic backgrounds were examined in >30 embryos.

Figure 2B Laminin timecourses second marker: nhr-67p::PM (JLF1402) n=6, PKC-3 (JLF1469) n=7, ceh-27p::LifeAct (JLF1391) n=9, NCAM-1 (JLF1378) n=6.

Figure S5D PAR-6^vir localization quantification is from PAR-6^vir single images: RFP::PKC-3 (JLF1440) n= 70 immobilized with CO₂, DLG-1::mScarlet-I (JLF1424) n=64 immobilized with CO₂, mScarlet-I::CAP-1 (JLF1798) n=38, immobilized due to high concentration of bacteria, PAR-6^int (JLF1776) n = 29, imaged before twitching, nhr-67p::PM/end-1p::PM (JLF1426) n=18 no immobilization. Observation of PAR-6^vir puncta appearance in 8 timecourses (PKC-3 n=4, nhr-67p::PM n=4) and puncta merging in 13 timecourses (PKC-3 n=2, CAP-1 n=7, nhr-67p::PM n=4).

Visualization, background subtraction, and image processing

‘Slice’ corresponds to single 0.3 or 0.5um optical imaging plane and was chosen to best visualize the cells/structures of interest. ‘Max’: Maximum intensity projections were taken through the entirety of the tissue of interest for visualizations. ‘3D’: 3D projections are screenshots of napari 3D rendering using mip (min/max intensity projection of inverted/greyscale images) with camera angle kept constant across time. 3D outlines of cell boundaries/interface were drawn manually on each z-stack before rendering as ‘iso-categorical’ for entire cells (Figure 3A, Figure S4A, Figure 4D) and ‘translucent’ to flatten the 3D model for laminin timecourse (Figure 2B) and interface in Figure 3A. For all timecourses, unless noted, 0’ is vir cell division (or within 5’ as estimated from tail spike progenitor divisions) and 1.5-fold stage is set to ~120’ post vir cell division. Videos were created from Cellular Morphology of C. elegans Embryo database³⁷, napari-animation, or Fiji distribution of imageJ, and subsequently annotated in Adobe Photoshop for final rendering.

Imaging parameters and linear brightness/contrast adjustments were consistent between all images compared across time and/or genotype and were applied to entire image with the following exceptions:

Figure 1A, Figure 4B, and Figure 4D colored by tissue: Multiple versions of PKC-3/PAR-6^vir (black) with colored PM channel (cyan-I, I-Forest, BOP orange-I, I-Bordeaux) channel made with same LUT values in Fiji were cropped to desired tissue in Adobe InDesign and spliced together. Unspliced single-channel images are shown in Figure S1A, Figure S5E, and Figure S5G.

Figure 2D, Video S2 : Application of ‘highlight’ to the image shown in Figure 2C. Intensities within a mask manually drawn along the apical surface were adjusted to a suitable level and added to original image in which nuclei were not oversaturated. This combined image was then rendered in 3D over time to visualize the merging of the apical surfaces along with the rectal nuclei. Note that gamma adjustment alone was not sufficient to allow for apical and nuclei visualization simultaneously and hence a ‘highlight’ was used in this case.

Figure S1C: Median filter (despeckle) applied to Figure S1C due to low expression of LET-656.

Figure 2F, S3C: Rolling-ball background subtraction (250 pixel radius) was applied to remove uneven background outside of the eggshell before contrast adjustments in order to show HMR-1/PKC-3 puncta merging.

Figure 3A,B, S4A, Video S3: See ‘cyst-to-tube’ transition below.

Figure S5L: gut granule autofluorescence was removed from the green CUTL-18 channel. A far-red 641nm excitation/emission image was taken along with the CUTL-18 and EPS-8 channels and all channels were background subtracted with 50 pixel rolling ball. The far-red autofluorescence image was transformed using 1.5-pixel opening (tophat transform) was applied and multiplied by 30 before using image calculator to subtract the resulting autofluorescence from the CUTL-18 image.

Figure S6H: Mitotic regulators example 1.5-fold embryos have the green intestinal end-1p::PM, elt-2p>N channel adjusted individually to enable visualization of intestinal morphology in each image. PKC-3 contrast adjustments were applied uniformly. In all images shown and in all quantifications, gamma=1.0 except as listed in figure legends in cases where both bright nuclei and dim apical/PM needed to be visualized in the same channel: vir ablation 3D projections (Figure 3C: JLF914 nhr-67p::H1 and PKC-3, gamma=0.5 magenta LUT imaged in 561nm channel,) and PLK-1^{E8 gut(−)} larvae (Figure S6K: JLF1527 and JLF560 end-1p::PM, elt-2p>N, gamma=0.25 forest-I LUT imaged in 488 channel).

Figure S6M: Median filter (despeckle) applied to lateral bean and comma images.

Quantification & Statistical Analysis

Timeline construction and cell-cell contact quantification:

Cell-cell contact was plotted from segmentation data³⁷ using a minutes post fertilization (mpf) offset to 4-cell stage from which all 8 segmented embryos were originally aligned in time (estimated to be 110’). Cell-cell contact between int8R and int8L examined to identify timepoints with segmentation errors that were removed from graphs. Division timepoints and cell-cell contact area data were taken from published repository and contact between desired pairs of data was summed. ‘Gastrulation ends’ was taken as the timepoint when contact area between ABplpapppp and ABprppppaa was greater than 0. Intestinal polarized was estimated from visual inspection of the dorsal midline for columnar shaped cells, int5 intercalation³⁵ start was decided from lateral views and end from dorsal views, valve bilateral cell shape change was decided from ventral views.

Lumen width estimates

Distances between plasma membranes were measured in EM micrograph slice 868 from the Mulcahy late L1 that shows the midpoint of the vir lumen. Measurement ends were placed on the cell plasma membrane at the tips of microvilli in the intestine or edge of the cell in vir and in the rectal gland/K cells, excluding the abundant apical ECM.

Cyst-to-tube transition

Live images were taken every 2 min to catch the dynamics of PKC-3. Unexpectedly, laser power changed only in timepoint 7 for some embryos during multipoint acquisition, and therefore t7 only was multiplied by 1.3 (the ratio of t7/t8 background values) before further analysis. Rolling ball background subtraction (250 pixel radius) was performed for both PKC-3 and pha-4p::PM channels and PKC-3 intensity was corrected using single exponential bleach correction¹⁰⁶. Intensity was summed over time within box placed at the posterior intestine encompassing the entire int8 interface and divided by intensity of initial timepoint. Apical length and interface length in YZ were estimated from PKC-3 and pha-4p::PM membrane signal respectively (average of 3 measurements) from resliced, maximum projection images and divided to give percentage coverage of interface. Data for each embryo were aligned to time point when valve and intestine form clear cell-cell contact by pha-4p::PM.

Apical ratio in vir ablation embryos

For each embryo with ventral oriented toward the coverslip, PKC-3 intensity at the level of the most ventral end of the int8 nuclei was divided by PKC-3 intensity at the more dorsal intestinal midline (int7/8 junction).

HMR-1^nhr-67(−) quantification

Images were background subtracted using mean of three estimates of the slide background near the embryo inside the eggshell. Two points were manually placed on the lateral membranes of visible vir cell using brightest HMR-1 pixels and/or CAAX signal and intensity was averaged over a 3-pixel-radius ball placed at each point. For plotting, to account for variability in exposure times used on different imaging dates, intensities for each imaging date were divided by the median of the intensities for array negative control embryos. Strong degradation threshold was set at one standard deviation below the mean of array negative HMR-1 intensities.

Colocalization of PAR-6^vir and PKC-3

A 6-8μm line along end of rectum or intestine in the plane of apical surface where PKC-3 was brightest was used to calculate linescan intensity values measured with imageJ function Plot Profile. A Gaussian fit of PAR-6^vir was used to calculate center and the distance in both PAR-6^vir and PKC-3 channels was offset. A mean for 12 embryos measured was calculated by first linearly interpolating the linescan values to a common set of distances before plotting the mean and 95% CI of intensities.

Puncta intensity quantification

Points were manually placed at the center of the vir-intestine and vir-rectum puncta. If no clear puncta were visible even after adjusting contrast to show dim signals, the brightest pixel near the center of the two gut nuclei was used. Fluorescence signal was averaged over a 3-pixel-radius ball placed at each point for each embryo and the mean of three background measurements inside the eggshell was subtracted. To calculate the ‘int-facing puncta’ percentage, the intensity at the intestine-facing puncta was divided by the sum of the intestine-facing and rectal-facing puncta for each embryo. For Figure 5C, ‘rectal-facing’ intensity was shown for clarity and is equivalent to 100 – ‘int-facing’ percentage. To plot the ‘total’, the sum of the intestine-facing and rectal-facing puncta for each embryo was divided by the mean of all control embryos and divided by 2. Individual points for each embryo plotted using swarmplot along with mean and standard deviation of each condition.

Junction/Lateral intensity ratio

Intensity was scored blind to genotype and with images in randomized order. Sliding window maximum projections of 5 z-stacks were examined for the brightest HMR-1::GFP signal along membranes. For each embryo, 2-pixel wide lines were drawn along the visible lateral (the dorsovental-oriented membrane between intestinal rings quantified on the side closest to the coverslip) and junctional membranes for at least 3 intestinal cells per embryo and averaged. Nearby average cytoplasmic signal was subtracted before taking the ratio of junctional to lateral signal.

Statistics

Statistics were performed using Excel or scipy.stats¹⁰⁷. The unit of biological replication was considered one embryo. For comparisons among multiple groups (Figure S1G), one-way ANOVA was used, and if significant, a post-hoc Tukey’s test for pairwise comparisons. For Figure 3F: Fisher’s Exact Test was performed between all control and all HMR-1^nhr-67(−) animals regardless of degradation strength. For comparisons between two conditions (Figure 3D, Figure 5C, 5G, Figure 6D, 6F, 6H), a either a two-tailed student’s T-test or a nonparametric Mann-WhitneyU test was performed when data did not meet the assumption of normality. Note that for Figure 6H, the HMR-1 alleles differ between the groups (cp21 HMR-1::GFP and wy1254 HMR-1::ZF::GFP, see Figure S7) with HMR-1::GFP being slightly brighter and so comparisons cannot be made across all four conditions.

Supplementary Material

Video S1

Video S1: vir cells contact the intestine during embryogenesis, related to Figure 2.

Cell membrane segmentations from Guan et al³⁷ over time from ventral and lateral view showing the lineage of vir (orange) and posterior intestine (yellow) within the context of the entire embryo through comma stage.

Video S2

Video S2: vir cells sit at the site of apical continuity between the intestine and rectum, related to Figure 2.

Puncta merging from 1.5 – 2-fold stage embryo shown in Figure 2D, showing the application of the apical highlight of PKC-3 signal.

Video S3

Video S3: The intestine undergoes a cyst-to-tube transition, related to Figure 3.

Movie of the cyst to tube transition in pre-bean/bean stage embryo from Figure 3A. Bottom views show 3D renderings of cells and/or interface at every other time point.

Methods

Methods S1: Details of strains, alleles, plasmids, primers, related to STAR Methods

Tabs contain information on C. elegans strains in each figure, alleles and/or transgenes generated by injections, plasmids, and oligonucleotides. Cites ^48,51,99

Highlights.

• Rectal valve (vir) cells are required for intestinal and rectal lumen continuity

• Posterior intestine transitions from a closed-off cyst to open tube upon vir contact

• Vir cells build bipolar apical surfaces abutting the intestine and rectum

• Vir cell bipolarity requires intestinal E-cadherin

Data and code availability

• Microscopy data for example images are available in data repository at doi: 10.17632/7n37yyb9vz.4 Further raw microscopy data reported in this paper will be shared by the lead contact upon request.

• Code is available in data repository at doi: 10.17632/7n37yyb9vz.4 and https://github.com/feldworms/vir_paper_Cote2026

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.