The utse is an H-shaped syncytial uterine cell that underlies the C. elegans uterus and opens along the central crossbar region during egg-laying to allow embryo passage (Ghosh and Sternberg 2014). The utse is flanked laterally on both sides of the animal by a string of epidermal seam cells that run the length of the body. The utse cell and seam cells are each encased in a BM, which become linked by a specialized matrix whose known functional components include hemicentin, fibulin, and type IV collagen (Figure 1A) (Gianakas, Keeley et al. 2023). The BM-BM tissue connection supports the uterus during egg-laying and disruption of the linkage results in uterine prolapse (Figure 1B and C) (Vogel and Hedgecock 2001). The utse and seam cells are in contact by the mid L4 larval stage (Gianakas, Keeley et al. 2023), but when the initial contact is established is not known.

The utse is formed from the fusion of one specialized uterine cell, the anchor cell, with eight adjacent uterine pi cell progeny, forming the syncytial nine-nuclei utse during the early L4 stage (Newman, White et al. 1996, Ghosh and Sternberg 2014). To determine when the utse first contacts the seam cells, we used an anchor cell-specific mCherry fluorophore marker (cdh-3p::mCherry::PLC8^PH; Figure 1D, arrow) (Hagedorn, Yashiro et al. 2009) that diffuses into the uterine pi cell progeny during fusion and simultaneously viewed a GFP seam cell marker (scmp::GFP::CAAX) (Chapman, Li et al. 2008). We found that the utse contacts the seam cells during the early L4 stage as the utse syncytium forms (Figure 1D, arrowheads; n=5/5 cases observed). Shortly after utse-seam contact, the BM-BM linkage is initiated at the mid L4 stage by the assembly of low levels of hemicentin, which is secreted by the utse and fibulin-1, which hemicentin recruits from the extracellular fluid. Fibulin-1 and hemicentin protect the utse-seam cell connection from mechanical forces arising from body movement and uterine and vulval muscle contractions prior to egg-laying (Gianakas et al. 2023). Hemicentin also promotes recruitment of type IV collagen from the extracellular fluid, which allows the utse-seam BM-BM connection to resist the strong mechanical forces from egg-laying (Gianakas et al. 2023) (Figure 1D). Together, these results show that the utse and seam cells are in contact prior to deposition of matrix components that strengthen the utse-seam BM-BM tissue connection.

As type IV collagen is critical to the BM-BM connection, we were next interested in understanding how receptors for type IV collagen regulate the utse-seam BM-BM linkage. The two main receptors for type IV collagen are integrin and the discoidin domain receptor (DDR) (Fidler, Boudko et al. 2018). C. elegans harbor two integrin heterodimers composed of different α chains with a shared ý chain: αINA-1/ ýPAT-3 and αPAT-2/ ýPAT-3 (Clay and Sherwood 2015). The integrin INA-1 is expressed in the utse and its loss leads to uterine prolapse, suggesting a possible role in utse-seam attachment (Morrissey, Keeley et al. 2014, Gianakas, Keeley et al. 2023). Whether DDRs function at the utse-seam connection, however, is unknown.

C. elegans have two DDR genes, ddr-1 and ddr-2, which are paralogs and orthologs to the two mammalian DDRs (Vogel, Abdulhussein et al. 2006). Like mammalian counterparts, the C. elegans DDRs are comprised of discoidin and discoidin-like domains, followed by a transmembrane domain and a kinase domain (Figure 2A). To determine if C. elegans DDRs play a role in utse-seam BM-BM linkage, we examined several ddr-1 and ddr-2 mutants (Figure 2A). We first screened for the Rup phenotype caused by uterine prolapse. Animals harboring the ddr-1(ok874) deletion allele, which eliminates the intracellular portion of DDR-1, and animals with the ddr-1(tm382) allele containing an early stop codon truncating the majority of the predicted protein (Unsoeld, Park et al. 2013), did not show a Rup phenotype (Figure 2A and B). In contrast, animals homozygous for the ddr-2(tm797) allele, which introduces an early stop codon and is thought to be a null or strong loss-of-function (Unsoeld, Park et al. 2013), exhibited a significant Rup/uterine prolapse defect. Animals carrying the ddr-2(ok574) allele, which deletes a portion of the intracellular kinase domain (Unsoeld, Park et al. 2013), also trended towards displaying a significant Rup phenotype (Figure 2A and B). We further generated a full-length ddr-2 deletion allele, ddr-2(qy64), and confirmed that complete loss of ddr-2 led to penetrant uterine prolapse (Figure 2A and B). Taken together, these results indicate that DDR-2 protects against uterine prolapse and may play a role in mediating utse-seam BM-BM attachment.

To determine if the Rup phenotype observed in ddr-2 mutants is a result of perturbations in the utse-seam BM-BM tissue connection, we examined utse-seam association from the late L3 through young adult stages in animals harboring the ddr-2(qy64) deletion allele (Figure 2C). We visualized utse-seam association using the utse and seam cell membrane markers cdh-3p::mCherry::PLC8^PH and scmp::GFP::CAAX, respectively. Animals homozygous for ddr-2(qy64) had normal utse-seam cell association at the early L4 stage; however, beginning at the mid L4 stage, gaps in the utse-seam attachment were detected in about 30% of animals. The frequency of gaps increased to ∼40% of animals by the mid-to-late L4 and persisted at a similar penetrance to young adulthood (Figure 2C and D). These observations suggest that DDR-2 functions around the mid L4 to late L4 stages to promote utse-seam attachment. Next, we wanted to determine if the utse-seam detachment observed in ddr-2(qy64) mutants led to uterine prolapse. Wild-type and ddr-2(qy64) animals were assessed at the L4 larval stage for utse-seam attachment defects and tracked to the 72-hour adult stage, where they were examined for the Rup phenotype. All wild-type L4 larvae had intact utse-seam attachment and did not rupture by 72-hour adulthood (n=8/8 animals). However, among nine ddr-2(qy64) mutant animals examined, four exhibited utse-seam detachments at the L4 stage, all of which underwent uterine prolapse (Rup) by adulthood. The remaining five ddr-2 mutant animals, which had an intact utse-seam linkage, did not display the Rup phenotype. Taken together, these results suggest that DDR-2 promotes utse-seam attachment during early stages of tissue linkage.

As utse-seam detachment in ddr-2 mutant animals was first observed at the mid L4 larval stage, we next sought to determine if DDR-2 was expressed in the utse, seam, or both cell types at this time. Between the mid L4 and late L4 stage, the utse-seam tissue connection forms into a tongue-and-groove association between the utse (tongue) and seam (groove) (Figure 2D and Figure 2—video 1). To resolve whether DDR2 is localized within the seam, utse, or both (Vogel and Hedgecock 2001), we used either a utse (cdh-3p::mCherry::PLC8^PH) or seam cell marker (scmp::2xmKate2::PLC8^PH), and imaged endogenously tagged DDR-2 protein (DDR-2::mNG) (Keeley, Hastie et al. 2020) in lateral confocal sections from the L4 to young adult stages (Figure 3A). We first examined the seam cells contacting the utse (Figure 3B). At the early L4 stage DDR-2::mNG was faintly visible at the surface of the seam cells and found in a few intracellular punctae (Figure 3B, n=10/10 animals). By the mid L4 stage, DDR-2 was predominantly found in punctae in the region of attachment with the utse. Levels of punctae increased until the late L4 stage and then declined sharply by the young adult (Figure 3B). Notably, at the late L4, when DDR-2 punctate localization was at its highest, we could not detect any DDR-2 signal at the cell surface (Figure 3B, n = 9/9). We further examined DDR-2::mNG in seam cells that were not contacting the utse and found that DDR-2 levels were uniformly low from the early L4 through the young adult stages (Figure 3—figure supplement 1A). We next investigated DDR-2 localization in the utse contacting the seam. Like the seam cells at the tissue connection site, DDR-2 in the utse was punctate and peaked in levels between the mid and late L4 stages and then declined rapidly (Figure 3C). We conclude that DDR-2 is present at high levels and localized within punctate structures in the seam and utse cells at the time when utse-seam defects are observed in ddr-2 mutant animals.

Receptor tyrosine kinases (RTKs), including DDRs, are activated upon ligand binding and then rapidly internalized into endocytic vesicles: first the early endosome (marked by RAB-5), followed by subsequent sorting for either lysosome-mediated degradation via late endosomes (marked by RAB-7), or recycled back to the cell surface in recycling endosomes (marked by RAB-11) (Goh and Sorkin 2013). In vertebrates, the exposure of cells to collagen results in the internalization of DDR1b into Rab5a positive early endosomes (Mihai, Chotani et al. 2009). Phosphorylation of DDR1 does not occur until the onset of endocytosis and recycling, suggesting that receptor activation occurs in endocytic vesicles (Mihai, Chotani et al. 2009, Fu, Valiathan et al. 2013). As C. elegans DDR-2 appeared to be functioning during the mid L4 to late L4 stages, when levels of DDR-2 were highest and present in vesicular structures, we sought to determine whether DDR-2 was internalized within endocytic vesicles. We examined localization of DDR-2::mNG punctae in relation to utse-expressed markers of early endosomes (cdh-3p::mCherry::rab-5), late endosomes (cdh-3p::mCherry::rab-7), and recycling endosomes (cdh-3p::mCherry::rab-11) (Sato, Norris et al. 2014). DDR-2 was co-localized with all three classes of endosomes in the utse, with the highest overlap with RAB-5 early endosomes (Figure 3—figure supplement 1B). In addition, we examined an early endosome marker expressed in seam cells (scmp::mKate2::rab-5) and observed DDR-2::mNG punctae co-localized with RAB-5 vesicles (Figure 3—figure supplement 1C). The presence of DDR-2 in endocytic vesicles within the utse and seam suggested that DDR-2 signaling may be required for utse-seam attachment. Consistent with this, animals harboring the ddr-2(ok574) mutant allele, where a portion of the intracellular kinase domain is deleted (Figure 2A), exhibited significant utse-seam detachments at the late L4 stage (Figure 4A). Taken together, these observations suggest that DDR-2 functions in the utse and seam cells to promote attachment at the tissue linkage site between the mid and late L4 stages, where it may signal within endocytic vesicles.

DDR signaling is triggered by collagen binding, and vertebrate DDRs bind multiple collagens with different binding affinities (Leitinger 2014). Genetic interaction studies in C. elegans revealed that transmembrane collagen COL-99, which is most similar to transmembrane collagen type XIII, XXIII, and XXV in vertebrates (Tu, Huhtala et al. 2015), may act as a ligand for C. elegans DDR-1 and DDR-2 during axon guidance and that type IV collagen might act as a ligand for DDR-2 during axon regeneration (Hisamoto, Nagamori et al. 2016, Taylor, Unsoeld et al. 2018). C. elegans BMs also harbor type XVIII collagen (CLE-1) (Keeley, Hastie et al. 2020). Thus, we next investigated whether any of these cell and BM-associated collagens could function as a ligand for DDR-2 at the utse-seam BM-BM tissue connection site.

Type IV collagen enriches at the utse-seam BM-BM linkage during its maturation and helps the connected tissues resist the mechanical forces of egg-laying. In contrast, type XVIII collagen is not present enriched at the linkage (Gianakas, Keeley et al. 2023). Whether COL-99 localizes to the utse-seam tissue linkage is unknown. We thus first examined an endogenously tagged COL-99 (mNG::COL-99) (Jayadev, Morais et al. 2022) and compared its localization and levels to type IV collagen (α1-type IV collagen::mNG (EMB-9::mNG)) and type XVIII collagen (CLE-1::mNG) (Keeley, Hastie et al. 2020) at the utse-seam connection in late L4 animals. COL-99 was not detected while EMB-9 was present at nearly 8-fold higher levels than CLE-1, suggesting that type IV collagen might be the ligand for DDR-2 (Figure 4B). Furthermore, the enrichment of type IV collagen levels at the utse-seam connection site occurs rapidly from the mid L4 to late L4 stage (Gianakas, Keeley et al. 2023), correlating with the buildup of DDR-2 in endosomes in the utse and seam cells.

As exposure to collagen triggers vesiculation of vertebrate DDR1b into endocytic vesicles, we wanted to determine whether type IV collagen controls vesiculation of DDR-2 at the utse-seam BM-BM connection. RNAi-mediated depletion of type IV collagen from the L1 through the late L4 larval stages resulted in a sharp decrease in vesicular DDR-2::mNG in the utse and seam cells at the linkage site, accompanied by enrichment of DDR-2 at cell surfaces (Figure 4C; Figure 4—figure supplement 1A). In contrast, reduction of COL-99 and type XVIII collagen did not significantly affect DDR-2 vesiculation (Figure 4C; Figure 4—figure supplement 1B and C). Depletion of hemicentin, which plays a crucial role in recruiting type IV collagen to the BM-BM linking matrix (Gianakas, Keeley et al. 2023), also resulted in a significant reduction in DDR-2 vesiculation (Figure 4D; Figure 4—figure supplement 1D). In addition, we did not detect DDR-2 at the cell surface, suggesting that hemicentin has a role in recruiting DDR-2 to the site of utse-seam attachment. Together, these findings indicate that type IV collagen assembly at the site of utse-seam BM-BM tissue linkage triggers DDR-2 vesiculation and may act as a ligand for DDR-2 activation to promote utse-seam attachment.

Since DDR-2 localizes to both the utse and seam cells, we next investigated where DDR-2 functions to promote utse-seam BM-BM attachment. We generated genome edited lines with single-copy insertions of dominant negative (DN) ddr-2 (Bernadskaya, Brahmbhatt et al. 2019) driven by the cdh-3 promoter for utse expression or the scm promoter for seam expression. Using the utse-specific membrane marker (nas-22p::2xmKate2::PLC8^PH) (Park, Pan et al. 2010) and a seam cell membrane marker (wrt-2p::GFP::PLC8^PH) (Wildwater, Sander et al. 2011), we found significant utse-seam detachments in both utse-expressed and seam-expressed DN ddr-2 animals at the late L4 stage (Figure 4F and G). Consistent with the timing and vesicular pattern of DDR-2 localization, these results indicate that DDR-2 functions in both the utse and seam between the mid and late L4 stages to promote utse-seam attachment.

The matrix components fibulin-1 and hemicentin play early roles in mediating utse-seam attachment and hemicentin promotes recruitment of type IV collagen and further enrichment of fibulin, which strengthens the BM-BM attachment (Gianakas et al. 2023). We thus next asked whether DDR-2 promotes utse-seam linkage by regulating the assembly of fibulin, hemicentin, or type IV collagen at the connecting matrix. Knockdown of ddr-2 by RNAi did not alter fibulin levels at the late L4 stage (Figure 5—figure supplement 1A and B). However, we observed a modest ∼20% reduction in hemicentin upon depletion of DDR-2 (Figure 5—figure supplement 1C). As loss of hemicentin results in utse-seam detachment at the late L4 stage (Gianakas, Keeley et al. 2023), we tested whether a ∼20% reduction of hemicentin could cause these defects by initiating RNAi at the early L4, just before the utse forms (Methods). We found that utse-seam attachments were intact in all animals at the late L4 stage upon ∼20% reduction of hemicentin (n=13/13 and 12/12 control and him-4 RNAi-treated animals examined respectively; Figure 5—figure supplement 1D), suggesting that the utse-seam detachment defect in ddr-2 mutant animals is not caused by a modest reduction in hemicentin levels. Interestingly, there was an ∼40% increase in type IV collagen levels at the utse-seam connection after RNAi mediated reduction of ddr-2, indicating a possible feedback mechanism between DDR-2 signaling and collagen assembly (Figure 5—figure supplement 1E). In sum, these results indicate that knockdown of ddr-2 does not decrease the functional levels of key matrix components at the site of utse-seam BM-BM connection, suggesting that DDR-2 does not promote utse-seam attachment by regulating the assembly of the BM-BM linking matrix.

Studies have indicated that vertebrate DDRs are not strong adhesive receptors (Xu, Bihan et al. 2012). Consistent with this, we detected weak DDR-2::mNG fluorescence signals at the cell membranes of the utse and seam. However, we observed strong DDR-2 signals in a vesicular pattern in both tissues. We thus assessed several molecular activities regulated by DDR signaling that might mediate utse-seam attachment. Vertebrate DDRs promote matrix metalloproteinase (MMP) expression and localization (Leitinger 2014). The C. elegans genome harbors six MMP genes, named zinc metalloproteinase 1-6 (zmp-1-6) (Altincicek, Fischer et al. 2010). We examined four available reporters of ZMP localization (ZMP-1::GFP, ZMP-2::GFP, ZMP-3::GFP, and ZMP-4::GFP) (Kelley, Chi et al. 2019). Only ZMP-4 was detected at the utse-seam connection and its localization was not altered by knockdown of ddr-2 (Figure 5—figure supplement 1F), suggesting that DDR-2 does not promote utse-seam linkage through regulation of MMPs.

DDRs also modulate integrin activity (Leitinger 2014). Cell culture studies have shown that vertebrate DDRs can enhance integrin adhesion and integrin cell surface levels in multiple cell types (Xu, Bihan et al. 2012, Staudinger, Spano et al. 2013, Bayer, Grither et al. 2019). C. elegans harbor only two integrin receptors, made up of either the α subunit INA-1 or PAT-2 dimerized to the sole β subunit PAT-3 (Clay and Sherwood 2015). We previously found that depletion of INA-1 results in a Rup phenotype, but the specific effect on utse-seam BM-BM tissue attachment was not determined (Morrissey, Keeley et al. 2014). Thus, we next asked whether DDR-2 promotes utse-seam attachment through regulation of integrin. First, we analyzed the localization and levels of endogenously tagged INA-1 (αINA-1::mNG) and PAT-2 (αPAT-2::2xmNG) (Jayadev, Chi et al. 2019) and found that both integrin α subunits were expressed in the utse and seam tissues and localized to the utse-seam connection region from the mid-L4 stage through young adulthood (Figure 5A and B; Figure 5—figure supplement 2A and B). Their levels mirrored that of DDR-2, with a peak at the late L4 stage, followed by a drop off in the young adult stage (Figure 5A and B). However, lNA-1 levels were at least sixfold greater than PAT-2 (note, each PAT-2 is linked to two mNG molecules, see Methods), suggesting it may have a dominant role in adhesion (Figure 5A and B). Loss of ddr-2 resulted in a significant reduction in levels of both INA-1 and PAT-2 at the utse-seam BM-BM attachment through the L4 larval stage (Figure 5B). Taken together these data suggest DDR-2 could function to promote integrin adhesion within each tissue at the utse-seam connection.

We focused on a possible role for DDR-2 in regulating INA-1 at the utse-seam linkage, as INA-1 was present at higher levels compared to PAT-2 (Figure 5A and B) and was more enriched at cell surfaces (Figure 5—figure supplement 2C). Furthermore, PAT-2 anchors muscle attachments (Moerman and Williams 2006, Gieseler, Qadota et al. 2017) (Figure 5A) and its loss leads to paralysis, which eliminates mechanical stress on the tissue connection and prevents defects in the utse-seam attachment (Gianakas, Keeley et al. 2023). To determine if DDR-2 promotes integrin activity, we reduced INA-1 levels at the late L4 stage to 30% by RNAi and assessed whether loss of ddr-2 enhanced the Rup and utse-seam splitting defects (Figure 5—figure supplement 2D and E). Loss of ddr-2 strongly enhanced the uterine prolapse defect caused by RNAi-mediated knockdown of ina-1 (Figure 5C; Figure 5—figure supplement 2D). In addition, loss of ddr-2 enhanced the frequency of utse-seam detachments at the late L4 stage upon depletion of INA-1 (Figure 5D). We note that utse-seam detachments were not detected upon RNAi against ina-1 alone, likely due to incomplete knockdown at this timepoint (Figure 5—figure supplement 2D and E). Together, these observations are consistent with a possible role for DDR-2 in regulating integrin function to promote utse-seam connection.

To determine where integrin functions to promote utse-seam attachment, we used a previously characterized dominant negative strategy, where expression of the ý integrin PAT-3 lacking the extracellular domain (referred to as DN integrin; Figure 5E) inhibits endogenous integrin function (Martin-Bermudo and Brown 1999, Lee, Cram et al. 2001, Hagedorn, Yashiro et al. 2009). We used the utse-specific zmp-1^mK50-51 promoter (Figure 5—figure supplement 2F) (Hagedorn, Yashiro et al. 2009) and the seam-specific wrt-2 promoter to disrupt integrin activity in the respective tissues through the time of utse-seam tissue linkage (late L3 to young adult stages, Figure 1D). Expression of DN integrin in either the utse or the seam cells resulted in utse-seam detachment at the late L4 (Figure 5F and G). Taken together, these findings indicate that DDR-2 may promote utse-seam BM-BM adhesion in part by regulating integrin levels and function.

As DDR-2 promotes integrin localization to the utse-seam BM-BM linkage and integrin activity promotes utse-seam attachment, we further explored how DDR-2 may regulate integrin. Talin and vinculin are core components of the integrin adhesion complex that support integrin activation (Chastney, Conway et al. 2021). We thus examined endogenously tagged talin (GFP::TLN-1) (Walser, Umbricht et al. 2017) and vinculin (DEB-1::mNG) localization and detected both proteins at the utse-seam attachment site. Levels of each protein ramped up to the late L4 and then decreased by the young adult stage, mirroring DDR-2 (Figure 6A and B). However, loss of ddr-2 only decreased DEB-1 levels (Figure 6B). Confirming a role in utse-seam connection, depletion of DEB-1 by RNAi resulted in utse-seam detachment at the late L4 stage (Figure 6C; Figure 6—figure supplement 1). These data show DDR-2 also regulates vinculin levels at the utse-seam linkage and that vinculin is required for utse-seam attachment.

Integrin adhesions form stable complexes that anchor cells to ECM (Chastney, Conway et al. 2021). To determine whether DDR-2 regulates integrin complex stability at the site of utse-seam BM-BM tissue connection, we analyzed the dynamics of INA-1 by performing Fluorescence Recovery After Photobleaching (FRAP) experiments. We focused on INA-1::mNG at the late L4 stage when DDR-2 peaks in levels. Strikingly, loss of ddr-2 nearly doubled the rate of recovery of INA-1::mNG fluorescence signal in the bleached region 10 minutes post-bleaching as compared to wild-type animals (Figure 6D). These observations indicate that DDR-2 stabilizes the INA-1 integrin adhesion complex at the utse-seam adhesion site.

Vertebrate DDR2 activates multiple signaling networks, including the PI3K/Akt and Ras/Raf/Erk cascades (Payne and Huang 2014, Chen, Kong et al. 2021). Notably, Ras regulates integrin activity in several mammalian cell types (Zhang, Vuori et al. 1996, Conklin, Ada-Nguema et al. 2010, Sandri, Caccavari et al. 2012, Lilja, Zacharchenko et al. 2017). We thus hypothesized that DDR-2 could promote integrin stabilization through Ras signaling. We first strongly depleted LET-60 (see Methods) the C. elegans Ras ortholog, and observed penetrant utse-seam attachment defects, which were not worsened in ddr-2 null mutant animals (Figure 7A). This suggests that DDR-2 and LET-60/Ras function in the same pathway to mediate the utse-seam connection. Next, we analyzed INA-1::mNG levels and stability after LET-60 reduction. RNAi against let-60 resulted in significantly lower INA-1 levels at the utse-seam BM-BM connection site (Figure 7B, pre-bleach panels and Figure 7—figure supplement 1). Further, like loss of ddr-2, FRAP analysis indicated that depletion of LET-60/Ras reduced the stability of INA-1 at the utse-seam linkage site (i.e., there was more rapid recovery of INA-1 after photobleaching (Figure 7B). We also generated animals harboring a gain-of-function mutation in Ras (LET-60^G13E, let-60(qy203)) (Singh and Han 1995). Strikingly, utse-seam detachments were observed in let-60 gain-of-function animals (Figure 7C) and FRAP experiments revealed that INA-1 had an approximately two-fold greater stability at the utse-seam BM-BM connection site (Figure 7D). Collectively, these findings suggest that Ras acts downstream of DDR-2 to set the appropriate levels and stability of INA-1 mediated adhesion during the formation the utse-seam BM-BM tissue connection.

C. elegans strains used in this study are listed in the Key resources table. Worms were reared on nematode growth medium plates seeded with OP50 Escherichia coli at 16°C, 18°C, or 20°C according to standard procedures (Stiernagle 2006). We used C. elegans vulval development to accurately stage animals through the morphogenesis of the utse-seam connection (late L3 to young adult stage) (Mok, Sternberg et al. 2015).

To generate the genome-edited mNeonGreen (mNG) knock-in allele for deb-1 (qy48), we used CRISPR/Cas9 genome editing with a self-excising cassette (SEC) for drug selection (hygromycin treatment) as described previously (Dickinson, Ward et al. 2013, Keeley, Hastie et al. 2020). An 18 amino acid flexible linker attached to the mNG fluorophore was inserted in frame and directly upstream of the stop codon. sgRNA sequences directing Cas9 cleavage near the C-terminus are provided in the Key resources table.

The ddr-2 knockout allele ddr-2(qy64), an 8615 bp deletion corresponding to the coding sequence of DDR-2, was also generated with the SEC CRISPR method detailed above with one modification—we generated a new starter repair plasmid lacking the fluorescent tag. Primers used to amplify homology arms, sgRNA sequences to direct Cas9 cleavage, and genotyping primers used to verify genome-edited knockout animals are provided in the Key resources table.

For the LET-60^G13E gain-of function mutation, let-60(qy203), we used the same strategy as the ddr-2 knockout allele. The point mutation was incorporated into the primer for homology arm amplification. Note that this mutation is identical to the let-60(n1046) allele.

For the utse marker qy91 [nas-22p::2xmKate2::pH], we generated a single-copy transgene inserted into the ttTi5605 transposon insertion on chromosome II (Frokjaer-Jensen, Davis et al. 2012) using the SEC CRISPR method detailed above. A 2 kb nas-22 promoter, two tandem mKate2 fluorophores attached to a PH domain, and an unc-54 3’ UTR fragment were ligated in order via Gibson assembly into the pAP087 starter repair plasmid containing homology arms for the ttTi5605 site. The sgRNA directing Cas9 cleavage near this region is contained in the pDD122 vector.

To express mKate2-tagged RAB-5 in the seam, we generated a single-copy transgene inserted into the ttTi4348 transposon insertion site on chromosome I (Frokjaer-Jensen, Davis et al. 2012) using SEC CRISPR. Briefly, the scm promoter, mKate2 fluorophore, rab-5 genomic sequence, and unc-54 3’ UTR were Gibson assembled in order into the pAP088 starter repair plasmid containing homology arms for the ttTi4348 site. The sgRNA directing Cas9 cleavage near this region is contained in the pCFJ352 vector.

For tissue-specific expression of dominant negative (DN) DDR-2 in the utse or seam, we also generated single-copy transgenes inserted into the ttTi4348 site. For utse expression, the following fragments were Gibson assembled in order into pAP088: a 1.5kb cdh-3 promoter fragment, DN ddr-2, mNG fluorophore, and unc-54 3’ UTR. Expression in transgenic animals was verified by assessing mNG fluorescence. For seam expression, the following fragments were Gibson assembled in order into pAP088: the scm promoter fragment, DN ddr-2, and mKate2 fluorophore. Expression in transgenic animals was verified by assessing mKate2 fluorescence.

The seam cell marker qyEx605 [scmp::2xmKate2::PH] was built by Gibson assembling a 1.3 kb scm promoter fragment, a fragment with two tandem mKate2 fluorophores attached to the PH domain, and unc-54 3’ UTR fragment in order. The construct was expressed as extrachromosomal arrays as described previously (Yochem and Herman 2003).

To build the construct for DN integrin expressed in the seam, we first amplified the DN integrin fragment from animals harboring the qyIs15 allele (see Key resources table). The DN integrin fragment is composed of a pes-10 enhancer element, the ost-1 signal peptide, the pat-3 ý-tail fragment, and unc-54 3’ UTR as described previously (Lee Cram et al. 2001). The DN integrin fragment was fused to a 1.3 kb wrt-2 promoter by PCR. The construct was expressed extrachromosomally.

All RNAi constructs except for ddr-2 RNAi were obtained from the Ahringer and Vidal RNAi libraries (Kamath and Ahringer 2003, Rual, Ceron et al. 2004) or generated previously (Gianakas, Keeley et al. 2023). For the ddr-2 clone, a ∼2.4 kb cDNA fragment corresponding to the longest ddr-2 transcript was amplified by PCR and inserted into the T444t vector as described previously (Gianakas et al 2023). RNAi experiments were performed using the feeding method (Timmons, Court et al. 2001) according to previously detailed protocols (Jayadev, Chi et al. 2019, Gianakas, Keeley et al. 2023). For experiments assessing the utse-seam connection, RNAi was initiated in synchronized L1 larvae and animals were examined at the late L4 stage (∼ 44 h treatment). For examination of uterine prolapse, RNAi was performed from the L1 through adulthood (∼ 120 h treatment). For depletion of hemicentin initiated at the early L4 stage, synchronized L1 worm were grown on control RNAi for ∼ 38 h to the early L4, and then transferred onto him-4 RNAi to the late L4 (∼5-6 h treatment). We verified knockdown efficiencies for all RNAi experiments except for let-60 RNAi by including a control with the relevant fluorescent tagged target protein, achieving between ∼75 to 100% reduction. We verified let-60 RNAi knockdown efficiency by plate level assessment of brood size, which was strongly reduced compared to control animals, as previously reported (Ceron, Rual et al. 2007).

All fluorescence images (except for photobleaching experiments) were acquired at 20°C on Zeiss Axio Imager A1 microscopes controlled by the µManager software v.1.4.23 or v.2.0.1 (Edelstein, Amodaj et al. 2010), equipped with a Hamamatsu ImagEM electron multiplying charge-coupled device camera or an Orca-Fusion scientific complementary metal oxide semiconductor camera, Zeiss 40x and 100x Plan Apochromat (1.4 numerical aperture) oil immersion objectives, Yokogawa CSU-10 spinning disc confocal scan heads, and 488 nm, 505 nm, and 561 nm laser lines. Worms were mounted on 5% noble agar pads containing 0.01M sodium azide for imaging.

For photobleaching experiments, images were acquired at 20°C on a Zeiss 880 single-point scanning confocal attached to a Zeiss Axio Observer Z1 microscope, with a Marzhauser linearly encoded stage, a 40x Plan Neofluar (1.3 numerical aperture) oil immersion objective, and a 488 nm laser line. Worms were anesthetized by soaking in 5 mM levamisole in M9 buffer for 15 min and then transferred to 4% noble agar pads. Coverslips were then placed and sealed on top with valap (equal weight Vaseline, lanolin, and paraffin) and flooded with 5 mM levamisole. Coverslip sealing kept the animals hydrated during the experimental timeframe, ensuring accurate imaging of endogenous protein dynamics (detailed in (Kelley, Wang et al. 2017)). A rectangle region within one arm of the BM-BM connection BLINK was photobleached using 30 iterations of simultaneous 405 nm and 488 nm excitation at 100% laser power for a total bleaching time of 1.5 seconds. Worms were imaged prior and immediately following bleaching, and then again 10 min later.

Brightfield images of worms were acquired at 20°C on a Zeiss Axio Zoom V16 stereo fluorescence microscope controlled by Zen 3.2 software, equipped with an Axiocam digital camera and a 3x objective. Worms were imaged without immobilization.

To image the utse and seam cells ventrally, we manually oriented animals in the ventral orientation (Kelley, Wang et al. 2017) and acquired z-stacks at 0.37 µm intervals, capturing the entire H-shaped structure of the utse and the seam cells flanking both sides of the utse at 100x magnification. Ventral views of the utse and seam shown in figures are maximum intensity projections of these z-stacks. 3D isosurface renderings of the utse and seam cells shown in Figure 2D were generated with Imaris 7.4 software (Bitplane). We acquired lateral z-stacks at 0.37 µm intervals to fully capture the superficial half of the utse-seam attachment.

For lateral imaging of DDR-2::mNG together with the utse, seam cells, or endosome markers expressed in the utse or seam, we acquired z-stacks at 0.37 µm intervals to capture as much of each tissue that was in focus superficially at 40x magnification. Images shown in figures are single slices where DDR-2 was sharply in focus within the tissue of interest. The same parameters were used for imaging INA-1::mNG and PAT-2::mNG in the utse and seam. Images in Figure 5A correspond to the z-slices where INA-1 and PAT-2 were most in focus at the BM-BM connection region. In Figure 5—figure supplement 2A and B, images shown are sum projections of the seam or utse regions.

To image DDR-2::mNG specifically at the utse-seam attachment site, we acquired lateral z-stacks at 0.37 µm intervals capturing both the surface of the utse and the seam cells. We used a utse cell marker as a reference to accurately capture this region (see also schematic in Figure 3A). Images shown in Figure 4 are sum projections of these z-stacks.

For lateral imaging of respective tagged proteins at the BM-BM connection region, we acquired single slices (40x magnification) at the middle focal plane where the BM-BM connection signal was sharply in focus. All images were process in Fiji 2.0 (Schindelin, Arganda-Carreras et al. 2012).

To assess utse-seam detachment, we used maximum intensity projections of the utse and seam viewed ventrally as detailed above. We defined utse-seam detachment as any gap in the utse-seam attachment of at least 4 square pixels in area. We preferred the ventral view over 3D rendering for this metric as gaps were variable in size and location (typically occurring in one or two arms of the utse on opposite sides). The ventral projections allowed us to reliably examine all four arms of the utse and whether they were in contact with the seam from animal to animal. The visual representation of a gap in a ventral maximum projection is an underestimate of the true size of the detachment, as the tongue-in-groove association of the utse and seam can only be seen in a 3D surface rendering of the utse and seam viewed from the side. We thus used a binary scoring system (animals with or without utse-seam detachments) instead of quantifying the size of the detached region. We note that ventral imaging of the utse and seam sometimes results in small utse-seam detachments as we manually orient the animals on the slide. To control for this, we acquired large datasets (at least 40 animals examined) for every control strain we used to visualize the utse-seam attachment with. We found that small detachments related to ventral imaging occurred in ∼8-12% of animals, depending on the genetic background. All datasets shown in figures have been normalized to account for this.

All quantifications of mean fluorescence intensity were done on raw images in Fiji 2.0. We drew ∼5-pixel long and 2-pixel wide linescans to obtain raw values of mean fluorescence intensity. For measurement within the BM-BM connection region, we positioned the linescan within the arm that was most in focus. Other regions where linescans were performed are indicated in figures. Background intensity values were obtained with similar linescans in adjacent regions with no visible fluorescence signal. Note that we used two tandem mNG fluorophores to visualize PAT-2 at the utse-seam attachment site (PAT-2::2xmNG) as it was present at very low levels. We imaged PAT-2::mNG and PAT-2::2xmNG in this region at the same exposure and found that the latter was approximately two times brighter (PAT-2::2xmNG mean fluorescence intensity 3941 ± 847 A.U., PAT-2::mNG 2037 ± 465 A.U., n = 10 animals examined each). Taken together with quantifications in Figure 5B, INA-1 is thus likely present at six-fold higher levels than PAT-2 at the utse-seam attachment site. For additional quantification related to FRAP, see below.

For FRAP analysis, we measured mean fluorescence intensity within the bleached region. In addition, we performed an equivalent linescan within the unbleached arm at the BM-BM connection to calculate a bleach correction factor to account for general photobleaching during image acquisition across the duration of the experiment, as previously described (Gianakas, Keeley et al. 2023). Briefly, the background-corrected fluorescence intensity measurement on the side of the BM-BM connection where FRAP was not performed at the 10 min post-bleach timepoint was divided by the respective value at the pre-bleach timepoint to obtain the bleach correction factor. Fluorescence intensity measurements at the pre-bleach and post-bleach timepoints on the side of the BM-BM connection where FRAP was performed were then multiplied by the respective bleach correction factor to normalize these values to the 10 min post-bleach fluorescence intensity measurement in the same region.

To count the number of DDR-2 punctae at the site of utse-seam attachment, we used sum projections of DDR-2::mNG capturing the surface of the utse and seam (described earlier). We restricted measurements to the utse-seam connection corresponding to one utse arm. Brightness and contrast was then adjusted to manually count individual punctae, as DDR-2 punctae showed a large variance in size and fluorescence intensity. Note that in the type IV collagen depletion condition, we only counted discrete puncta and did not include the large cell surface accumulation of DDR-2 in the analysis.

To determine co-localization of DDR-2 with markers of endocytic vesicles, we also used a manual approach due to the large variability of the size and fluorescence intensity of punctae. We selected a single slice where the utse or the seam was sharply in focus, and restricted measurements to a single utse arm or a region of the seam corresponding to a single utse arm. For each animal, the number, and positions of DDR-2 punctae were noted and compared to the respective endosome marker (RAB-5, RAB-7, or RAB-11). If a DDR-2 puncta physically overlapped with or was in contact with an endosome punctum, it was classified as one that co-localized with the endosomal marker. We then calculated the percentage overlap of DDR-2 punctae with the vesicle marker for each animal (corresponding to each data point within boxplots in Figure 3—figure supplement 1B and C).

Synchronized L1 larvae were plated (∼20 animals per plate) and after 24 h, the exact number of worms on each plate was recorded. Plates were then visually screened for ruptured worms (uterine prolapse) every 24 h, up to 120 h post L1 plating. Ruptured animals were scored and removed from plates to avoid double counting and after 120 h, the percentage of animals with uterine prolapse was calculated. At least 50 animals were screened for every experiment.

For schematics in Figure 1A and Figure 8, model protein sizing was scaled according to previously generated models (Keeley, Hastie et al. 2020). We sized hemicentin, fibulin, type IV collagen, integrin, and DDR-2. We also approximated the span of the basement membranes of the utse and seam cells (∼50 nm each), as well as the BM-BM connecting matrix (∼200 nm), based on electron microscopy of the utse-seam attachment (Vogel and Hedgecock 2001).

Statistical analysis was performed in GraphPad Prism 9. Normality of datasets was assessed using the D’Agostino-Pearson normality test. We used parametric tests for datasets that followed a Gaussian distribution and non-parametric tests for those that did not. For comparisons of means between two populations, we used either an unpaired two-tailed Student’s t test or a Mann-Whitney U test. For comparisons of means between three or more populations, we performed either a one-way ANOVA followed by post hoc Dunnett’s test or Kruskal-Wallis H test with post hoc Dunn’s test. For comparisons between two categorical variables, we used the Fisher’s exact test. All graphs were prepared in GraphPad Prism. Figure legends indicate sample sizes, statistical tests used, and P values.