Lmo7 recruits myosin II heavy chain to regulate actomyosin contractility and apical domain size in Xenopus ectoderm

Apical constriction, or reduction in size of the apical domain, underlies a variety of morphogenetic processes during embryonic development (Martin and Goldstein, 2014; Perez-Vale and Peifer, 2020; Sawyer et al., 2010). One consequence of apical constriction is the cell delamination required for epithelial-to-mesenchymal transition (EMT) (Ramkumar et al., 2016; Williams et al., 2012). Another consequence is tissue invagination that leads to epithelial furrow or tube formation (Eiraku et al., 2011; Sherrard et al., 2010; Wallingford et al., 2013). Significantly, failure of apical constriction results in epithelial morphogenesis defects, such as abnormal neural tube closure (Balashova et al., 2017; Brown and Garcia-Garcia, 2018; Colas and Schoenwolf, 2001; Haigo et al., 2003; Itoh et al., 2014; Kowalczyk et al., 2021).

The actomyosin system that drives apical constriction is a highly dynamic network of filamentous actin (F-actin) and non-muscle myosin II (NMII), crosslinked by α-actinin (Murphy and Young, 2015; Vicente-Manzanares et al., 2009). NMII heavy chains assemble tail-to-tail into bipolar filaments that form NMII mini-filaments (Bresnick, 1999; Heissler and Sellers, 2014; Shutova and Svitkina, 2018). The head domains of NMII mini-filaments bind F-actin and the sliding movements between F-actin and NMII generate contractile forces (Lecuit et al., 2011; Martin and Goldstein, 2014; Salbreux et al., 2012). Actomyosin networks can be arranged in many patterns, from loose mesh-like networks to highly organized sarcomere-like actomyosin bundles (Agarwal and Zaidel-Bar, 2019; Schwayer et al., 2016).

Two models have been proposed to explain how actomyosin networks regulate apical constriction in epithelial cells. The ‘purse-string’ model involves circumferential actomyosin bundles that run parallel to apical junctions (Boller et al., 1985; Coravos and Martin, 2016; Gorfinkiel and Blanchard, 2011; Hirano et al., 1987; Munjal et al., 2015; Yonemura et al., 1995). These bundles are linked to apical junctions by F-actin-binding proteins, including αE-catenin, vinculin and afadin at adherens junctions (AJs), and zonula occludens proteins (ZO-1-3) at tight junctions (TJs) (Abe and Takeichi, 2008; Drees et al., 2005; Fanning et al., 1998; Itoh et al., 1997; Sawyer et al., 2009; Yamada et al., 2005). NMII motor activity pulls antiparallel F-actin arrays and reduces the distance between sarcomere-like repeats. The other mechanism involves medioapical actomyosin meshwork that is also linked to apical junctions but generates pulling forces perpendicular to these junctions (Azevedo et al., 2011; Blanchard et al., 2010; David et al., 2010; Martin et al., 2009; Mason et al., 2013). In both models, the association of actomyosin network and apical junctions is essential for apical constriction, because the uncoupling leads to failure of apical constriction in canoe/afadin-deficient Drosophila embryos (Roh-Johnson et al., 2012; Sawyer et al., 2011, 2009).

To gain further insights into the interaction of the actomyosin network with apical junctional machinery, we initiated the study of LIM domain only protein 7 (Lmo7) that was originally identified as a binding partner of afadin and α-actinin (Ooshio et al., 2004). Although mice lacking Lmo7 are viable and develop relatively minor defects (Du et al., 2019; Lao et al., 2015; Mull et al., 2015; Tanaka-Okamoto et al., 2009), Lmo7 has been implicated in skeletal and cardiac muscle differentiation, and in metastatic cancer (He et al., 2018; Holaska et al., 2006; Liu et al., 2021; Nakamura et al., 2011; Ott et al., 2008). Notably, the Drosophila homologue Smallish is essential for epithelial morphogenesis and controls apical domain size in the follicular epithelium (Beati et al., 2018). These observations suggest that Lmo7 may have uncharacterized functions in the reorganization of actomyosin and apical junctions in vertebrate epithelia.

This study describes Lmo7 as a novel regulator of actomyosin contractility and apical constriction in Xenopus ectoderm cells. This activity of Lmo7 requires the Rho-ROCK-NMII system and the primary function of Lmo7 appears to be the binding and recruitment of NMII heavy chain to the actomyosin networks in the apical region of cells. Lmo7 knockdown delays neural tube closure in Xenopus embryos. We propose that Lmo7 is a scaffold protein that regulates crosstalk between apical junctions and the actomyosin system.

To understand Lmo7 function in epithelial morphogenesis, we studied Lmo7 protein distribution in Xenopus ectoderm, a tissue that exhibits dynamic remodeling of apical junctions (Higashi et al., 2016). In stage 11 ectoderm, Lmo7 localized at cell-cell boundaries (Fig. 1A). Ectoderm cryosections confirmed Lmo7 apicolateral distribution to the area corresponding to apical junctions (arrowheads in Fig. 1C,C’). This distribution was different from the localization of other components of apical junctions. First, Lmo7 was visible as two distinct parallel bands along cell-cell boundaries, especially in short cell contact areas (Fig. 1B). By contrast, ZO-1, a tight junction (TJ) component, and β-catenin, a adherens junction (AJ) component (Hartsock and Nelson, 2008), were detected as a single band flanked by two Lmo7 bands (Fig. 1D-H′). Second, Lmo7 accumulated near but not at tricellular junctions (Fig. 1B′). These observations establish the localization of Lmo7 in the proximity of apical junctions. In addition to apical junctions, Lmo7 puncta were visible at the apical cortex (arrowheads in Fig. 1F,F″).

We next asked whether Lmo7 localization at the cell-cell boundaries corresponds to the circumferential actomyosin bundles that run parallel to apical junctions (Yonemura et al., 1995). The major components of actomyosin bundles are F-actin, NMII and α-actinin. We observed that both exogenous GFP-NMIIA (Fig. 1I′,J′) and endogenous NMIIA (Fig. 1L,M,N′,O′) form two bands along the cell-cell boundaries. Lmo7 was positioned in a way largely overlapping with the NMIIA bands (Fig. 1I-J″,N-O″), but somewhat closer to the plasma membrane at both bicellular (Fig. 1J-J″,O-O″,R) and tricellular (Fig. 1K-K″) junctions. In addition, Lmo7 double line closely overlapped with phalloidin staining (Fig. 1P-R). These findings document Lmo7 localization at peri-junctional actomyosin bundles (Fig. 1S).

A previous study showed that Smallish, the Lmo7 homologue in Drosophila, caused apical domain reduction in follicular epithelial cells (Beati et al., 2018). We therefore asked whether Lmo7 has a similar activity. Injection of GFP-Lmo7 RNA triggered hyperpigmentation and invagination of ectoderm at the site of injection (Fig. 2A,B). Because pigment granule accumulation is commonly associated with apical F-actin bundle formation in Xenopus ectoderm (Kurth and Hausen, 2000), increased pigmentation serves as a good indicator of apical constriction (Haigo et al., 2003; Itoh et al., 2014; Merriam et al., 1983; Morita et al., 2010; Ossipova et al., 2014; Popov et al., 2018). In addition, we confirmed apical constriction by measuring apical domains en face (Fig. 2C,D) and by assessing the ratio of apical width to basolateral length of cells in cross-sections (Fig. 2E-G) after phalloidin co-staining. To deplete Lmo7, a translation-blocking morpholino oligonucleotide (lmo7-ATGMO) was injected in the ectoderm. Although no overt effect on ectoderm pigmentation was apparent, lmo7 knockdown expanded the apical domain (Fig. 2H-I). This effect of lmo7MO was rescued by co-injection of GFP-Lmo7 RNA (Fig. S1). This is consistent with the Lmo7 function in apical domain size regulation.

A highly organized actomyosin network is known to generate the contractile force required for apical constriction. We compared peri-junctional actomyosin bundles in Lmo7-expressing and control non-expressing cells (Fig. 3). Lmo7-expressing cells exhibited thick double bands of Flag-Lmo7 in a sarcomere-like pattern (Fig. 3A,C). The effect was less pronounced in embryos injected with lower dose of lmo7 RNA (Fig. 1E,J), suggesting dose-dependent effects. Flag-Lmo7 induced the formation of intense F-actin double bands that largely overlapped with Lmo7 (Figs 3A-D and 1P-Q″). The appearance of thick F-actin bundles was accompanied by the accumulation of phosphorylated myosin regulatory light chain (pMRLC) (Fig. 3E-H), NMIIA (Fig. 3I-L) and exogenous α-actinin 4 (Fig. S2). These results suggest that Lmo7 promotes actomyosin bundle formation at apical junctions.

We next examined Lmo7 effects on the medioapical actomyosin network that may also contribute to apical constriction in Xenopus ectoderm (Arnold et al., 2019). A small fraction of Lmo7 localized to the medioapical cortex and formed puncta (Fig. 4A,C). These GFP-Lmo7 puncta colocalized with puncta of F-actin (Fig. 4C-D″) and NMIIA (Fig. 4E-F″). More apical puncta formed in cells expressing high amounts of Lmo7, indicating dose dependency (Fig. 4A-D″,G). Taken together, these results suggest that Lmo7 increases the density of the actomyosin network both at apical junctions and the medioapical cortex.

Rho-ROCK-NMII signaling has been shown to play a crucial role in apical constriction. The main outcome of this pathway is the phosphorylation of myosin regulatory light chains (MRLCs) by Rho-dependent protein kinase (ROCK) (Das et al., 2014; Haigo et al., 2003; Hildebrand, 2005; Hildebrand and Soriano, 1999; Itoh et al., 2014; Nakajima and Tanoue, 2011; Nakajima and Tanoue, 2012; Nishimura and Takeichi, 2008; Popov et al., 2018). To test whether the Lmo7 effect on Xenopus ectoderm requires the Rho-ROCK-NMII pathway, we inhibited MRLC phosphorylation by the co-expression of a dominant-interfering form of ROCK (ROK-C) or constitutively active myosin phosphatase Mypt1 (Mypt1CA) (Feng et al., 1999; Marlow et al., 2002; Weiser et al., 2009). Apical constriction was assessed by pigment granule accumulation (Fig. 5A-D) and the quantification of apical domain size (Fig. 5E-G). Co-expression of either RNA suppressed Lmo7-mediated apical constriction (Fig. 5C,D,F,G), indicating that Rho-ROCK-NMII pathway is required for this process.

The phosphorylation of MRLC promotes NMII bipolar filament formation and NMII binding to anti-parallel F-actin, and may initiate force generation. We therefore asked whether Lmo7-mediated actomyosin reorganization increases mechanical tension at apical junctions and/or the apical cortex. We took advantage of Ajuba family proteins that change their localization in response to tension (Chu et al., 2018; Ibar et al., 2018; Rauskolb et al., 2014; Razzell et al., 2018). One of the Ajuba family proteins, Wilms tumor 1-interacting protein (Wtip), forms aggregates or puncta. These puncta disappear in cells expressing Shroom3, a known inducer of apical constriction, potentially reflecting the increased tension at the cell cortex (Chu et al., 2018). We injected Flag-Lmo7 RNA in a subpopulation of RFP-HA-Wtip-expressing cells (Fig. 5H). The RFP-HA-Wtip puncta both at the apical cortex (Fig. 5L) and apical junctions (Fig. 5M) were no longer visible in Lmo7-expressing cells when compared with the non-expressing control cells (Fig. 5I-M). Whereas mechanisms and physiological significance of Wtip puncta formation warrant further investigation, these results suggest that Lmo7 increases tension in overexpressing cells.

Lmo7 has the calponin homology domain (CH), the domain of unknown function DUF4757, the α-actinin-binding region (Ooshio et al., 2004), the PDZ domain, two coiled-coil domains and the LIM domain that binds to afadin (Ooshio et al., 2004) (Fig. 6A). To define the domains of Lmo7 required for apical constriction, various Lmo7 constructs were expressed in the superficial ectoderm (Fig. S3). We found that the removal of the C-terminal LIM domain (Fig. 6B), the coiled-coil domain and the PDZ domain (Fig. 6C), or the N-terminal CH domain (Fig. 6F) did not affect the pigment granule accumulation. On the other hand, the α-actinin-binding region was necessary for this activity (Fig. 6D,E). DUF4757(aa 242-400) was also necessary (Fig. 6G) but not sufficient on its own (Fig. 6I). Lmo7(aa 242-709) containing both DUF4757 and the α-actinin-binding region was the shortest fragment that triggered apical pigment accumulation (Fig. 6H). We confirmed that Lmo7(aa 242-709), but not Lmo7(aa 242-400), reduced apical domain size in individual cells (Fig. 6K-M). These studies suggest that the Lmo7(aa 242-709) was both required and sufficient for apical constriction.

We next evaluated whether Lmo7 interacts with NMII heavy chain as reported for its paralog human Limch1 (Lin et al., 2017). Both endogenous NMIIA and NMIIB were found in Flag-Lmo7 pull-downs in transfected HEK293T cells (Fig. 7A). As DUF4757(aa 242-400) was required and sufficient for this binding (Fig. 7A), we will refer to it as the NMII-binding domain (NMIIBD). These results suggest that Lmo7 effect on actomyosin may be due to the NMIIBD interaction with NMII heavy chains. Consistent with our hypothesis, alanine substitution of the conserved WQ-WK amino acids in NMIIBD (Fig. 7B) eliminated the binding of NMIIA and IIB in HEK293T cells (Fig. S4A). Flag-Lmo7(AAAA) localized at apical junctions (Fig. 8A,B) but did not increase NMIIA accumulation at apical junctions (Fig. 8B,B′,H) as seen in wild-type Flag-Lmo7 (Fig. 8C,C′,G). Importantly, neither GFP-Lmo7(AAAA) nor GFP-Lmo7(AAAK) induced apical constriction in ectodermal cells (Fig. S4B-E). These results indicate that Lmo7 binding to NMII via NMIIBD is required for Lmo7-mediated apical constriction.

Although necessary, Lmo7 binding to NMII was not sufficient for apical constriction. Although Lmo7(aa 242-400) bound to NMIIA (Fig. 7A), its overexpression did not stimulate NMIIA accumulation at apical junctions (Fig. S5G,H) or induce apical constriction (Fig. 6). By contrast, Lmo7(aa 242-709) that contains an α-actinin-binding region was still able to stimulate NMIIA accumulation (Fig. 8D-F″,I) and triggered apical constriction (Fig. 6). Lmo7(aa 242-400) and Lmo7(aa 242-709) showed similar levels of accumulation at apical junctions (Fig. 5A-F″), suggesting that loss of apical constriction by Lmo7(aa 242-400) is not due to reduced localization at apical junctions. Notably, although full-length Lmo7 promoted the formation of thick and dense F-actin networks (Fig. 4), Lmo7(aa 242-709) did not exhibit these effects (Fig. S5I,J). These results suggest that NMII is a primary cellular target of Lmo7 in apical constriction, although other proteins, such as α-actinin, are involved.

The role of Lmo7 in NMII localization has been further confirmed in loss-of-function experiments. We observed reduced GFP-NMIIA localization at apical junctions in the cells depleted of Lmo7 (Fig. 9A-C). Nevertheless, the Lmo7-depleted cells preserved normal morphology with unaltered distribution of ZO-1 (Fig. 9D,E). F-actin accumulation at apical junctions was not altered either (Fig. 9F,G). Taken together, our findings suggest that Lmo7 functions to maintain NMII at apical junctions.

To evaluate Lmo7 function during different developmental processes, we first examined the temporal and spatial expression of Lmo7 transcripts in Xenopus embryos. RT-PCR analysis demonstrated ubiquitous presence of Lmo7 transcripts at different stages of embryonic development, including the maternal expression (Fig. S6A). In situ hybridization analysis revealed broad expression of Lmo7 in epidermal ectoderm and neuroectoderm (Fig. S6B-E). Of note, Lmo7 transcripts showed a ‘salt-and-pepper’ pattern in late gastrula dorsal ectoderm (Fig. S6B). In the late neurula, Lmo7 was detectable at high levels in preplacodal ectoderm (Fig. S6H,I), somitic mesoderm (Fig. S6I,J) and the surface neuroepithelial layer (Fig. S6F,G). In tailbud and more advanced embryos, Lmo7 RNA was detected in the skin and pharyngeal arches (Fig. S6K-O). Lmo7 was also expressed in the somites (Fig. S6K,L,N) and the heart (Fig. S6M), consistent with previously proposed role of Lmo7 in muscle development (Holaska et al., 2006; Lao et al., 2015; Mull et al., 2015).

Neuroectodermal expression of Lmo7 and its ability to induce apical constriction suggested a possible role of Lmo7 in neural tube closure. A translation blocking morpholino (MO) was injected into one side of embryos at 2- to 4-cell stage, together with gfp RNA to label the injected side. The morphants exhibited incomplete neural tube closure in stage 20 embryos (compare Fig. 9H and I; black arrow in Fig. 9I, Movie 1), indicating inefficient apical constriction. This modest defect in neural tube closure was quantified by measuring the width of neural tube opening in embryos injected with Lmo7 translation- or splicing-blocking MOs (Fig. 9J-L, Fig. S7C). Taken together, these results demonstrate the requirement of Lmo7 for neural tube closure in Xenopus embryos.

In this study, we identify Lmo7 as a novel regulator of actomyosin contractility and apical domain size in Xenopus embryonic ectoderm. Lmo7 localizes to actomyosin networks at apical junctions and the apical cortex. Embryonic cells expressing Lmo7 form thicker and denser actomyosin network when compared with control ectoderm. We find that Lmo7 binds NMII heavy chains via NMIIBD, a domain with previously unknown function, and observe a tight correlation between apical constriction inducing activity of Lmo7 deletion mutants and NMII recruitment to the cell cortex. We propose that Lmo7 triggers apical constriction by promoting NMII incorporation into actomyosin networks.

The striking localization of Lmo7 to two cortical bands on both sides of the apical cell-cell junction is similar to the distribution of NMIIA in the normal tissue, including non-sensory epithelial cells of the mammalian inner ear (Ebrahim et al., 2013). In gain-of-function experiments, Lmo7 promoted the formation of organized peri-junctional actomyosin bundles. Our knockdown experiments revealed expanded apical domain in ectoderm cells and reduced NMIIA at apical junctions. These observations suggest that Lmo7 triggers apical constriction via a ‘purse-string’ mechanism. Similar actomyosin enrichment with ‘sarcomere-like’ organization of NMII was described in MDCK cells depleted of ZO proteins, suggesting a crosstalk between apical junctions and actomyosin (Choi et al., 2016; Fanning et al., 2012; Tokuda et al., 2014; Van Itallie et al., 2009; Yu-Kemp et al., 2022). In contrast to ZO protein, which may be inhibitory, Lmo7 appears to promote the sarcomere-like organization.

The ability of Lmo7 to promote peri-junctional actomyosin bundles does not exclude a role for Lmo7 in the medioapical actomyosin network. The importance of medioapical actomyosin for apical constriction has been intensively studied in Drosophila and other invertebrates (An et al., 2017; Azevedo et al., 2011; Blanchard et al., 2010; David et al., 2010; Martin et al., 2009; Mason et al., 2013; Royou et al., 2002). We found increased junctional and medioapical F-actin and NMIIA in Lmo7-expressing cells. These observations lead us to propose that Lmo7 is responsible for the organization and contractility of apical actomyosin networks. Further studies will be necessary to understand how circumferential actomyosin and medioapical actomyosin cooperate to induce apical constriction in vertebrate epithelia and how Lmo7 involves in this process.

Compared with strong effects of Lmo7 overexpression in actomyosin reorganization, Lmo7 knockdown in Xenopus embryos showed relatively mild defects. The observed effects on normal ectoderm are consistent with a general function of Lmo7 in actomyosin network architecture and apical domain regulation, rather than exclusively in apical constriction. Nevertheless, in agreement with the role in apical constriction, Lmo7 morphants exhibited a delay in neural tube closure. The neural tube defects were relatively mild, suggesting that the loss of Lmo7 is functionally compensated for by other proteins, such as its paralogue Limch1 (Lin et al., 2017). Lmo7 knockout mice also exhibit limited developmental defects in specific epithelia (Du et al., 2019; Lao et al., 2015; Tanaka-Okamoto et al., 2009), indicating similar compensation mechanisms in higher vertebrates. On the other hand, Drosophila embryos lacking Smallish, the fly orthologue of Lmo7, are defective in epithelial morphogenesis (Beati et al., 2018), pointing to the indispensable role of Smallish in actomyosin control. Double knockout of Lmo7 and Limch1 is necessary to properly evaluate their loss-of-function phenotypes in vertebrates.

Our experiments show that the ability of Lmo7 to trigger apical constriction depends on the recruitment of NMII heavy chains. It remains unclear how Lmo7 increases the actomyosin contractions after NMII recruitment. Lmo7 may recruit or activate Rho or ROCK to increase MRLC phosphorylation. Alternatively, Lmo7 may stimulate ATPase activity of the NMII head motor domain, or increase NMII heavy chain phosphorylation (Conti et al., 1991; Dulyaninova et al., 2007; Even-Faitelson and Ravid, 2006; Murakami et al., 1998) or NMII interaction with other proteins (Dahan et al., 2012; Dulyaninova et al., 2005; Elliott et al., 2012; Garrett et al., 2006). Of note, NMIIBD is disrupted in certain splicing variants of Lmo7 (Fig. S8) and only the second half of it is conserved in Drosophila Smallish (Fig. 7B). In addition to recruiting NMII to apical junctions, Lmo7 might act as a scaffold to position F-actin, NMII and α-actinin in the circumferential actomyosin bundles, and organize them in a sarcomere-like pattern. The domains that may interact with F-actin are the CH domain and the afadin-binding LIM domain (Ooshio et al., 2004). Surprisingly, Lmo7(aa 242-709) lacking both the CH and LIM domains retained the ability to trigger apical constriction but did not cause any further visible F-actin enrichment. Because Lmo7(aa 242-709) contains domains that bind NMII and α-actinin, Lmo7(aa 242-709) may work as a linker to stabilize NMII and α-actinin in apical actomyosin networks. This would be consistent with the ‘ratchet’ model of apical constriction that includes repeated cycles of actomyosin contraction and stabilization (Martin et al., 2009). Further studies are needed to understand how Lmo7 promotes F-actin enrichment or polymerization in the apical domain of cells, and how Lmo7(aa 242-709) promotes apical constriction without such F-actin enrichment.

The effect of Lmo7 on actomyosin network is likely relevant to mechanotransduction processes at apical junctions. Lmo7 levels increase in response to the mechanical force (Hashimoto et al., 2019). Lmo7 binds afadin, which has been implicated in force transmission at cell-cell junctions (Choi et al., 2016; Manning et al., 2019; Yu and Zallen, 2020). Supporting this hypothesis, we found that Lmo7 influences the distribution of Wtip, a sensor of mechanical tension (Chu et al., 2018). The remaining challenge is to determine how the interplay of Lmo7, afadin and other junctional proteins leads to effective generation, sensing and transmission of mechanical forces during epithelial morphogenesis.

Xenopus laevis Lmo7 cDNA clone was obtained from Dharmacon. Full-length and truncated forms of Lmo7 were PCR amplified and cloned into a pCS107-Flag, -HA or -GFP expression vector using standard protocols. Primers used for the amplification are in Table S1. pCS2-RFP-HA-Wtip has been previously described (Chu et al., 2018). pCS2-GFP-NMIIA and pCS2-GFP-NMIIB were subcloned from CMV-GFP-NMHC II-A and CMV-GFP-NMHC II-B (Addgene 11347 and 11348)(Wei and Adelstein, 2000). pCS2-GFP-α-actinin 4 was constructed from pCS2-hiActTS-GR (a gift from Tatsuo Michiue, The University of Tokyo, Japan). pCS2-ROK-C and pCS2-Mypt1CA were from Florence Marlow (Icahn School of Medicine at Mount Sinai, New York, USA; Marlow et al., 2002) and David Kimelman (The University of Washington, Seattle, WA, USA; Weiser et al., 2009), respectively.

Wild-type Xenopus laevis adults were purchased from Nasco, and maintained and handled according to the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. A protocol for animal use was approved by the Institutional Animal Care and Use Committee (IACUC) at Icahn School of Medicine at Mount Sinai. In vitro fertilization and embryo culture were performed as described previously (Dollar et al., 2005). Embryo staging was determined according to Nieuwkoop and Faber (1967). For microinjections, embryos were transferred into 3% Ficoll 400 (Pharmacia) in 0.5× Marc's modified Ringer's (MMR) solution [50 mM NaCl, 1 mM KCl, 1 mM CaCl2, 0.5 mM MgCl2 and 2.5 mM HEPES (pH 7.4)] (Peng, 1991). Linearized plasmid DNA was used to synthesize capped RNA using mMessage mMachine SP6 Transcription kit (Invitrogen). Synthesized RNA was purified by LiCl precipitation. RNA in 5-10 nl of RNase-free water (Invitrogen) was microinjected into one or two blastomeres of four- to 16-cell stage embryos. Morpholino oligonucleotide (MO) specific for Lmo7 was purchased from Gene Tools, and had the following sequence: Lmo7-ATGMO, 5′-GAATTTTCATTCCATTCCATTG-3′; Lmo7-spMO, 5′-AAATGCAAAGAATGTACTTACTCGC-3′. GFP or memGFP RNA was co-injected with MO for lineage tracing. Imaging of GFP-negative cells was carried out at least 20 cells away from the boundary region, to take account of the difference in diffusion of RNA and MO.

HEK293T cells were purchased from ATCC and maintained in Dulbecco's modified eagle media (DMEM) (Corning) supplemented with 10% fetal bovine serum (Sigma) in a 37°C incubator with 5% CO₂. HEK293T cells grown in 60 mm dishes to 70-80% density were transfected with a mixture of 4 µg plasmid DNA and 20 µg polyethylenimine (Polysciences). Cell lysates were extracted 24 h after transfection.

Cell lysates were extracted from HEK293T cells using RIPA buffer [50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate and 0.1% SDS] supplemented with Protease Inhibitor Cocktail III (Calbiochem) and PhosStop phosphatase inhibitor cocktail (Roche). Cell lysates were incubated with anti-Flag M2 agarose (Sigma) at 4°C for at least 6 h. Agarose beads were washed three times in Tris-buffered saline containing 0.05% Triton X-100 (TBST). Agarose beads were heated at 95°C for 5 min in SDS sample buffer containing 5% β-mercaptoethanol (Sigma). After SDS-PAGE and transfer to a nitrocellulose membrane with a 0.2 µm pore size (BioRad), immunoblotting was carried out with the following antibodies: rabbit anti-NMIIA pAb (BioLegend, POLY19098, 1:500), rabbit anti-NMIIB pAb (BioLegend, POLY19099, 1:500), mouse anti-DYKDDDDK mAb clone 2H8 (Cosmo Bio, KAL-K0602, 1:1000) and mouse anti-GFP mAb clone B2 (Santa Cruz Biotechnology, sc-9996, 1:500). HRP-conjugated secondary antibodies against mouse or rabbit IgG were from Cell Signaling Technology (7074 and 7076, 1:5000). Chemiluminescent signals were acquired using Clarity ECL Western Blotting Substrates (BioRad) on the ChemiDoc MP Imaging System (BioRad).

Xenopus embryo lysates were made by incubating 20 embryos in 400 µl lysis buffer [50 mM Tris-HCl (pH 7.6), 50 mM NaCl, 1 mM EDTA and 1% Triton X-100] supplemented with protease inhibitor and protein phosphatase inhibitor cocktails as described above. After centrifugation for 3 min at 16,000 g, the supernatant was mixed with SDS-sample buffer, followed by standard SDS-PAGE and western blot protocol. Experiments were repeated at least three times.

Digoxigenin-11-UTP-labeled Lmo7L RNA probes, sense and antisense, were prepared from linearized plasmids containing Xenopus Lmo7 L cDNA, using SP6 or T7 RNA polymerase (Promega) and DIG RNA labeling kit (Roche). Three independent probes were used to confirm the specificity of Lmo7 L expression. In situ hybridization was conducted following a standard protocol described by Harland (1991), with some modifications. After hybridization, embryos were incubated with anti-digoxigen-AP antibody (Roche). BCIP/NBT Alkaline phosphatase substrate kit (Vector laboratory) was used for color reaction. Embryo images were captured using a Leica stereo-microscope equipped with a color CCD camera.

Capped RNAs encoding Lmo7, Lmo7 mutants, ROK-C or Mypt1-CA were injected into two animal-ventral blastomeres of four-cell embryos. For analysis of pigment granule accumulation, embryo images were captured using a Leica stereo microscope equipped with a color CCD camera. Each experiment included 20-30 embryos per condition. Experiments were repeated at least three times.

For formaldehyde fixation, Xenopus embryos were devitellinized manually using forceps and fixed in MEMFA [0.1 M MOPS (pH 7.4), 2 mM EGTA, 1 mM MgSO4 and 3.7% formaldehyde] (Harland, 1991) for 60 min at room temperature. After washing in phosphate-buffered saline (PBS), embryos were permeabilized in 0.1% Triton in PBS for 30 min. For trichloroacetic acid (TCA) fixation, devitellinized embryos were fixed in ice-cold 10% TCA for 30 min. Embryos were then incubated in 1% BSA in TBS at 4°C overnight for blocking. After that, embryos were incubated with primary antibodies in 1% BSA, 1% DMSO in PBS or TBS at 4°C overnight. Antibodies used in this study are the following: rabbit anti-NMIIA pAb (BioLegend, POLY19098, 1:300), rabbit anti-Myosin light chain (phospho S20) pAb (Abcam, ab2480, 1:500), mouse anti-DYKDDDDK mAb clone 2H8 (Cosmo Bio, KAL-K0602, 1:500), rat anti-HA mAb clone 3F10 (Roche, 118673230002, 1:100), mouse anti-ZO1 mAb clone T8-754 (Itoh et al., 1993), rabbit anti-β-catenin pAb (Santa Cruz Biotechnology, sc-7199, 1:500) and chicken anti-GFP pAb (Abcam, ab13970, 1:500). After washing in PBS or TBS three times, embryos were incubated with secondary antibodies conjugated with either Alexa 488, Cy3 or Cy2 fluorescent dyes in 1% BSA and 1% DMSO in PBS at 4°C overnight. Alexa 488-, Cy3- or Cy2-conjugated secondary antibodies against mouse, rat and rabbit IgG or chicken IgY were obtained from Invitrogen (A-32723 and A-11034, 1:300) or Jackson ImmunoResearch Laboratory (711-165-152, 715-165-151, 712-165-153 and 703-225-155, 1:500). AlexaFluor 555-phalloidin (Invitrogen, A34055) was added to the secondary antibody solution to label F-actin. After washing in PBS or TBS three times, embryos were transferred in 25% glycerol in PBS. The animal pole side of embryos was dissected manually by forceps and mounted on slide glasses using Vectashield mounting medium (Vector). Imaging was carried out using Zeiss AxioImager upright microscope with Zeiss Axiocam CCD camera or Zeiss LSM880 confocal microscope system. Images were captured from at least five independent embryos. Experiments were repeated at least three times. Captured images were processed and quantified using ImageJ software.

Lmo7 ATGMO (30 ng) was injected into two right blastomeres of four-cell stage embryos, with GFP RNA (100 pg) as a tracer. The vitelline membrane was manually removed using forceps at stage 15. An embryo was placed into a small hole on 1% agarose gel in 0.1×MMR to stabilize its position. Time-lapse imaging was carried out using the AxioZoomV16 fluorescence stereomicroscope (Zeiss) equipped with the AxioCam 506 mono CCD camera (Zeiss).

To quantify apical domain size, embryos injected with relevant RNA or MOs were co-stained with Alexa Fluor 555-conjugated phalloidin (ThermoFisher Scientific). To measure the apical domain surface area, animal pole ectoderm samples were flat-mounted en face on glass slides. After image acquisition, the apical domains of individual cells were manually outlined using a free-hand line tool in ImageJ based on phalloidin staining, and the area of the apical domain was measured using ImageJ. For Lmo7 cells, only cells that shared at least one boundary with control wild-type cells were used for quantification. For control cells, only cells that shared at least one boundary with Lmo7 cells were used.

To quantify the ratio of apical width to basolateral length, gelatin-embedded embryos stained with Alexa Fluor 555-conjugated phalloidin (ThermoFisher Scientific, A34055, 1:200) were cryosectioned as described previously (Ossipova and Sokol, 2021) using MEMFA as a fixative. After image acquisition, the apical membrane and the basolateral membrane of individual cells were manually outlined on cross-sections based on phalloidin staining. Their length was measured using ImageJ.

To quantify the enrichment of Lmo7, F-actin and endogenous NMIIA at apical junctions, five to ten circles ∼2 µm in diameter were randomly drawn in cell-cell boundaries. Only cell-cell boundaries formed between two Lmo7-expressing cells or two control cells were selected. Tricellular vertices were excluded for simplicity. To quantify medioapical enrichment of Lmo7, F-actin and NMIIA, a 10 µm circle was drawn in the center of individual cells. ImageJ was used to measure the integrated fluorescence intensity within the selected area.

To quantify the junctional enrichment of exogenously expressed GFP-NMIIA in Lmo7 morphants, individual cell-cell boundaries were marked with a 2 µm wide line, using a free-hand line tool in ImageJ. Tricellular vertices were excluded. The fluorescence intensity of GFP-NMIIA in these lines were measured and normalized to that of membrane-tethered RFP. To quantify Wtip puncta formation near apical junctions, cell-cell boundaries formed between Lmo7- and Wtip-expressing cells or between the cells expressing only Wtip (control cells) were selected. The fluorescence intensity of RFP-HA-Wtip per µm² was measured using ImageJ.

For statistical analysis, the Mann–Whitney U-test was used to determine the statistical significance between two groups. Statistical significance of the difference among more than three groups was assessed by Dunn's test using a Bonferroni correction for the P-values.

Neural tube opening was assessed in embryos at stage 16-18. A series of z-stack images was captured on the AxioZoomV16 fluorescence stereomicroscope (Zeiss) equipped with the AxioCam 506 mono CCD camera (Zeiss). Individual z-stage images were processed into a single image using the ‘extended depth of focus’ module on the Zeiss Zen (blue edition) imaging software. The widest distance from the midline to the fusing edge of the neural fold was measured using ImageJ.

We thank Robert Adelstein, Florence Marlow, Tatsuo Michiue and David Kimelman for plasmids, Pamela Mancini for help with Lmo7 mutagenesis at the early stages of this project, and members of the Sokol laboratory for discussions. We acknowledge the help from the ISMMS Microscopy Core facility.

The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.200236