Female adult X. laevis frogs were induced to ovulate after injection of human chorionic gonadotropin. Eggs were fertilized in vitro, after sacrificing male frogs and acquisition of testes. The embryos were dejellied in 2% cysteine (pH 7.8) and subsequently reared in .1 × Marc’s Modified Ringers (MMR). All experiments and experimental protocols were approved by the National Committee for the Protection of Animals Used for Scientific Purposes of the Republic of Cyprus with the license (CY/EXP/PR.L05/2022).

For microinjections, embryos were placed in a solution of 4% Ficoll in .33 × MMR and injected using a glass capillary pulled needle, forceps, a Singer Instruments MK1 micromanipulator and Harvard Apparatus pressure injector at the 4-cell stage according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1994). After injections, embryos were transferred for 1 h in 4% Ficoll in .33 × MMR and then washed and kept in .1 × MMR. Injected embryos were allowed to develop to neurula stages (Nieuwkoop and Faber stage 12-5–13) and imaged live or allowed to develop to the appropriate stage and the dissected or fixed in 1x MEMFA for 1–2 h at room temperature. Capped mRNAs encoding fluorescent protein fusions were in vitro transcribed using mMessage machine kits (Ambion). The amount of mRNA per 4 nl of microinjection volume was as follows: membrane-GFP, 80 pg; Utrophin-GFP, 80 pg, FERM-FRNK (FF)-GFP: 500 pg.

The Shroom3 (Haigo et al., 2003), Vangl2 (Darken et al., 2002) and Wee2 MO (Leise and Mueller, 2004) morpholinos (MO) were previously described and were ordered from GeneTools. 30 ng of Shroom3 MO, 20 ng of Vangl2 MO, 30 ng Wee2 MO were injected per blastomere.

The amount of MO used was based on what was used in previous studies and typically close to what was shown to elicit maximum downregulation efficiency. Specifically, 20 ng Vangl2 MO was previously used and validated via immunofluorescence (Darken et al., 2002; Ossipova et al., 2015). Furthermore, this morpholino at this concentration phenocopies convergent extension defects elicited by dominant negative constructs for PCP proteins (Butler and Wallingford, 2018; Christodoulou and Skourides, 2022). The Shroom3 morpholino has been validated previously and successfully phenocopies the phenotypes induced by a Shroom3 dominant negative construct (Haigo et al., 2003; Lee et al., 2007; Chung et al., 2010). The amount of Shroom3 morpholino previously used for neural plate targeted injections is the same as the amount used in this study (30 ng) (Inoue et al., 2016). The Wee2 morpholino has been validated previously and the amount used in our study a (30 ng) is within the range used previously for efficient Wee2 downregulation (Leise and Mueller, 2004). Shroom3 and Vangl2 MOs were injected at the 2 dorsal blastomeres of 8 cell stage embryos to target the neural plate. Wee2 MO was targeted at the somitic mesoderm by targeted injections laterally of the dorsal marginal zone.

Immunofluorescence was performed as previously described (Zanardelli et al., 2013). Briefly, embryos were fixed for 2 h at room temperature, permeabilized in PBST (1 × PBS, .5% Triton, 1% dimethyl sulfoxide) and blocked for 1 h in 10% donkey serum. Primary antibodies were incubated overnight at 4°C. We used a primary antibody against integrin MyoD hybridoma supernatant (1:50, D7F2, Hybridoma Bank). Embryos were washed in PBST and incubated for 2 h with secondary antibodies at RT, washed several times and post-fixed in 1XMEMFA. The secondary antibodies used were Alexa fluor 488 (1:500, A21202, Invitrogen), Phalloidin was incubated together with the secondary antibodies, phalloidin 546 (1:500, A22283, Invitrogen).

Live imaging of neurula stage Xenopus embryos was performed on a ZEISS LSM 710 confocal microscope. The ZEISS ZEN software was used during imaging. Embryos were imaged in a custom chamber made of thick layer of vacuum grease on a microscope slide and sealed with a coverslip. The embryos were imaged in their vitelline membrane, and the height of the custom-made chamber was optimized for each embryo to ensure minimal contact with the coverslip thus achieving immobilization while preventing flattening of the embryos and minimizing external mechanical force application. Embryos were mounted in 0.1 × MMR and kept at room temperature during imaging.

All image analysis and quantification were carried out using Fiji software (Schindelin et al., 2012). Single cell tracking in Figure 3F was carried out using the manual tracking plugin of Fiji. For somitic mesoderm convergent extension index in Figures 1–3 the dorsoventral length of the presomitic mesoderm was divided by its mediolateral length. Somite boundaries were defined either by cell shape (Somitic mesoderm cells: columnar elongated shape, Lateral plate mesoderm cells: cuboidal shape) for Figures 1, 2 or MyoD expression for Figure 3. The neural plate midline in Figures 3C, 4D, 5D and neural plate boundary in Figure 3C, 4D, were determined by retrograde tracking. The neural plate boundary in Figure 7 was determined based on the elevated F-actin levels in the neural plate compared to those of the surface ectoderm.

GraphPad Prism 8.0 software was used for all statistical analysis performed. The sample size of the experiments carried out was defined based on previous experimental experience. Quantitative data presented, shows the mean ± S.E.M, or the total number of datapoints obtained. The statistical tests carried out on the quantitative data obtained are annotated in each figure legend.

We have recently shown that the surface ectoderm is mechanically coupled with the neural plate and its development is permissive for neural tube closure in Xenopus embryos (Christodoulou and Skourides, 2022). Specifically, our recent work revealed that the medial movement of the surface ectoderm during neural tube closure is passive and driven by the remodeling of the neural plate. Additionally, we have shown that increase surface ectoderm tension results in defective neural tube closure. During neural tube closure the neuroepithelium is also mechanically coupled with the underlying somitic mesoderm through the ECM which contains fibronectin, collagen IV, laminin and proteoglycans (Davidson et al., 2004; Mole et al., 2020). Additionally during this stage of development the somitic mesoderm undergoes convergent extension (Wilson et al., 1989) which results in its anteroposterior lengthening, mediolateral thinning and dorsoventral expansion (Leise and Mueller, 2004). At the same time the neural plate lengthens along its anteroposterior axis, thins along the mediolateral axis and bends to form the neural tube (Christodoulou and Skourides, 2015; Christodoulou and Skourides, 2022). We hypothesize that in the event that mesodermal morphogenesis is required for neural plate morphogenesis the evolution of tissue shape changes will be temporally linked. To test this hypothesis, we quantified the width of the neural plate and the degree of somitic mesoderm convergent extension in a mixed population of stage 13 (early neurula) to stage 16 (late neurula) embryos. This analysis revealed that neural tube closure and somitic mesoderm convergent extension are indeed temporally linked (Figures 1A, B). Specifically, neural tube closure progression, evident by the reduction of neural plate width, is accompanied by concomitant dorsoventral expansion and mediolateral thinning of the somitic mesoderm (Figures 1A, B). The tight temporal correlation of the two processes and taking into account the direct interaction of the neuroepithelium with the somitic mesoderm (Davidson et al., 2004), suggest that morphogenetic remodeling in both tissues might result in a reciprocal mechanical communication which will directly impact tissue morphogenesis in either tissue. Such mechanical control of tissue morphogenesis has been described before in other model systems (Butler et al., 2009; Christodoulou et al., 2019; Xiong et al., 2020). Specifically, during Drosophila development, mouse early post-implantation development and chick embryo axis elongation, the behavior of a tissue has a direct impact on its neighboring tissue morphogenesis through mechanical coupling.

Surface ectoderm medial movement at neurula stages is passive and relies on neural plate morphogenesis (Christodoulou and Skourides, 2022). Thus, based on the above data we asked if somitic mesoderm convergent extension is affected by neural plate morphogenesis. To address this possibility, we disrupted neural tube closure by abrogating either convergent extension or apical constriction. To impair neural plate convergent extension, we downregulated a known regulator of convergent extension, Vangl2 (Devenport, 2014), by targeted microinjections of Vangl2 morpholino (Darken et al., 2002). Specifically, to target the neuroepithelium we injected Vangl2 MO at the 2 dorsal/animal blastomeres of 8 cell stage embryos. In order to disrupt apical constriction, we unilaterally downregulated Shroom3 within the neuroepithelium by targeted injections of Shroom3 morpholino (Haigo et al., 2003). As expected, knockdown of both Vangl2 and Shroom3 resulted in impaired convergent extension and apical constriction respectively, as well as defective neural tube closure (Figures 2A, C). Comparison of somitic mesoderm convergent extension in control and Vangl2 morphant embryos, as well as in non-injected and Shroom3 morphant neural plate sides, revealed that somitic mesoderm convergent extension occurs normally both in the absence of neural plate convergent extension and apical constriction (Figures 2A–D). Overall, these data show that despite the direct interaction between the two tissues, somitic mesoderm morphogenesis takes place independently from the neural plate and is not impacted by defects in either neural convergent extension or apical constriction during neural tube closure.

The somitic mesoderm also directly interacts with the surface ectoderm during primary neurulation. During mouse embryogenesis, after completion of neural tube closure paraxial mesoderm morphogenesis depends on the presence of the surface ectoderm (Correia and Conlon, 2000). Therefore, we went on to examine if surface ectoderm development affects somitic mesoderm convergent extension during neural tube closure stages. For this, we used a previously characterized dominant negative construct for FAK, FERM-FRNK (FF), which has been shown to lead to thickening and stiffening of the ectoderm epithelium during gastrula and neurula stages (Petridou et al., 2013; Christodoulou and Skourides, 2022). As expected, unilateral targeted injections of FF-GFP led to thickening of the surface ectoderm. This was accompanied by defects in neural tube closure (Figure 2E), as previously described (Christodoulou and Skourides, 2022). In contrast to the effect on neural plate, SE thickening did not affect somitic mesoderm convergent extension during neural tube closure stages (Figure 2F). These data indicate that somitic mesoderm convergent extension occurs independently of surface ectoderm expansion.

Overall, the above data demonstrate that somitic mesoderm convergent extension is an active process and occurs independent from both neural plate and surface ectoderm morphogenesis. Importantly, these data are in agreement with the fact that somitic mesoderm explants can undergo convergent extension in the absence of the neuroepithelium and the surface ectoderm (Wilson et al., 1989). Our data further show that in the presence of neural plate and surface ectoderm defects somitic mesoderm morphogenesis is unaffected. The latter is in agreement with data demonstrating that the paraxial mesoderm is twice as stiff as the neuroepithelium at neurula stages (Zhou et al., 2009). This implies that mechanical resistive forces generated by a deformed neuroepithelium, or surface ectoderm are not able to impede morphogenesis of the stiffer somitic mesoderm which can generate forces that can overcome softer tissue resistance.

Neither neural plate nor surface ectoderm defects can appreciably impact somitic mesoderm morphogenesis likely due to the higher stiffness and resulting elevated force generation capacity of the somitic mesoderm compared to that of adjacent tissues. This would presumably allow somitic mesoderm morphogenesis to elicit passive deformation of adjacent tissues moderating potential defects from arising in the stiffer tissue when surrounding softer tissues that are not properly formed. Based on this interpretation, it would be expected that defects in the remodeling of the somitic mesoderm would mechanically impact the morphogenesis of the neuroepithelium. We went on to examine the impact of somitic mesoderm morphogenesis on neural tube closure. To accomplish this, we decided to specifically disrupt somitic mesoderm convergent extension during Xenopus neurulation. The tyrosine kinase Wee2 has been previously shown to display tissue restricted expression within the paraxial mesoderm (Leise and Mueller, 2004). Importantly, Wee2 is responsible for the transient cell cycle arrest within this tissue, which is necessary for somitic mesoderm convergent extension (Leise and Mueller, 2004). Therefore, Wee2 is an ideal target to specifically disrupt somitic mesoderm convergent extension during neurulation. We proceeded with targeted unilateral microinjection of a previously characterized morpholino against Wee2 (Leise and Mueller, 2004), targeting the paraxial mesoderm (Supplementary Figures S1A, B). Upon Wee2 downregulation, Wee2 morphant somitic mesoderm failed to thin along the mediolateral axis and expand along the dorsoventral embryo axis (Figures 3A, B), showing that Wee2 downregulation results in impaired somitic mesoderm convergent extension, as previously described (Leise and Mueller, 2004).

Having established an effective approach to specifically to disrupt somitic mesoderm convergent extension, we went on to examine the impact of defective somitic mesoderm convergent extension on neural tube closure. We performed live imaging of Wee2 morphant embryos and analyzed fixed morphant embryos. These experiments revealed that unilateral defects in somitic mesoderm convergent extension resulted in defective neural tube closure (Figures 3C–E; Supplementary Figure S1C; Supplementary Movie S1). Specifically, analysis of neural plate width in fixed embryos revealed that the NP was wider at the side injected with Wee2 MO (Figure 3D; Supplementary Figure S1C). Furthermore, time lapse recording revealed slower movement of neural folds towards the midline of the neural plate resulted in wider NP at the Wee2 morphant side (Figures 3E–G; Supplementary Movie S2). Overall, these data show that somitic mesoderm convergent extension is necessary for neural tube closure. Our data are in agreement with recent studies showing that paraxial mesoderm morphogenesis has a positive impact on neural tube closure during mouse embryogenesis (de Goederen et al., 2022; Li et al., 2022; Nychyk et al., 2022). However, what aspects of neural plate morphogenesis are affected by defects in somitic mesoderm development is not known.

Hinge points form at the lateral edges of the neural plate and are visible as areas with increased actomyosin accumulation (Haigo et al., 2003; Christodoulou and Skourides, 2015). Examination of paraxial mesoderm Wee2 morphant neurula stage embryos revealed that F-actin accumulation at the edge of the neural plate overlying the Wee2 morphant somitic mesoderm was abolished (Figures 4A–C). Based on these observations we conclude that hinge point formation is defective when somitic mesoderm morphogenesis is impaired. F-actin cable establishment and maintenance are both sensitive to tension forces and actin filaments have been suggested to directly sense mechanical force (Sun et al., 2020; Winkelman et al., 2020). Hingepoints, are cable like structures which are under constant mechanical stimulation, since are located at the boundary of the actively remodeling neural plate. We decided to examine if somitic mesoderm convergent extension is needed for the initial formation or is necessary for the subsequent maintenance of the hinge point. To accomplish this, we performed live imaging using Wee2 morphant embryos, injected with Utrophin-GFP, an F-actin marker compatible with live imaging (Christodoulou and Skourides, 2015; Christodoulou and Skourides, 2022). Time lapse-recording revealed that the hinge point, indicated by the reduction of cell surface area and subsequent accumulation of medioapical F-actin, fails to form at the neural plate side overlying Wee2 morphant somitic mesoderm (Figure 4D; Supplementary Movie S3). This shows that somitic mesoderm morphogenesis is necessary for the initial formation of hinge points during neural tube closure.

Hinge points form upon dramatic change of the shape of neural plate lateral boundary cells (Haigo et al., 2003; Christodoulou and Skourides, 2015). This cell shape change of boundary cells is a consequence of apical constriction (Christodoulou and Skourides, 2015). The defects in hinge point formation when somitic mesoderm convergent extension is impaired, suggest that apical constriction within the neuroepithelium requires proper somitic mesoderm shaping. The hallmark of apical constriction is the reduction of apical cell surface area (Sawyer et al., 2010; Martin and Goldstein, 2014). Quantification of the apical surface area of neuroepithelial cells from control or morphant somitic mesoderm in fixed embryos revealed that the apical surface of neuroepithelial cells overlying Wee2 morphant somitic mesoderm was larger (Figures 5A-A’, B). Additionally, time lapse recordings of embryos expressing Utr-GFP and have unilateral Wee2 morphant somitic mesoderm revealed that neuroepithelial cells at the morphant side fail to reduce their apical surface area (Figures 5C, D; Supplementary Movie S4). These data show, as expected based on the lack of hinge point formation, that apical constriction within the neuroepithelium is defective when somitic mesoderm convergent extension is impaired. Importantly examination of neighbor exchanges, a cell behavior necessary for neural plate convergent extension (Blankenship et al., 2006; Butler and Wallingford, 2018; Christodoulou and Skourides, 2022), revealed that this behavior is unaffected in Wee2 morphants (Supplementary Figure S2; Supplementary Movie S5), suggesting that somitic mesoderm convergent extension specifically impacts neuroepithelial cell apical constriction. The latter is partially supported by data showing that deep neural tissue explants autonomously undergo convergent extension (Elul et al., 1997).

Apical constriction during neural tube closure is driven by calcium mediated actomyosin contraction pulses (Christodoulou and Skourides, 2015). The stepwise reduction of the apical cell surface is maintained by a ratchet mechanism which prevents cell relaxation between each contraction step (Martin et al., 2009; Krueger et al., 2020). We and others have shown that the contraction and stabilization steps during apical constriction have distinct regulators (Martin et al., 2009; Christodoulou and Skourides, 2015). To examine which step of apical constriction is affected by defects in somitic mesoderm morphogenesis we analyzed time lapse recordings of embryos expressing the F-actin marker Utrophin-GFP. Our analysis revealed pulsed accumulation of medioapical F-actin in neuroepithelial cells overlying control somitic mesoderm. Similarly, the pulsed accumulation of medioapical actin was present in neuroepithelial cells overlying Wee2-morphant somitic mesoderm (Figures 6A,A’,B,B’; Supplementary Figures S3A–C; Supplementary Movie S6, S7). As shown medioapical actin accumulation is followed by a reduction of the apical cell surface area in neuroepithelial cells overlying normal somitic mesoderm. In contrast, accumulation of medioapical actin is not followed by efficient reduction of the apical surface in cells overlying morphant somitic mesoderm (Figures 6C,D; Supplementary Figure S3D; Supplementary Movie S6, S7). This phenotype is reminiscent of defects in apical ratcheting during apical constriction (Krueger et al., 2020). These findings suggest that while the genetic program necessary for the initiation of contraction pulses during neuroepithelial cell apical constriction is unaffected, tensile forces exerted on the neuroepithelium by the mispositioned underlying somitic mesoderm affect the stabilization of the actomyosin machinery necessary for ratcheted apical constriction, possibly through a mechanosensitive signaling pathway. Importantly, the molecular regulators of apical constriction have been shown to be mechanosensitive (Pouille et al., 2009). Thus, our data suggest that increased resistive forces stemming from defective somitic mesoderm morphogenesis impact on mechanosensitive molecules regulating actomyosin contractility during apical constriction.

Collectively, our data indicate that somitic mesoderm morphogenesis is necessary for neural tube closure. They also show that defects in somitic mesoderm convergent extension directly impact intrinsic neural plate morphogenesis, specifically apical constriction. Given the above, we postulated that mild defects in neural plate and somitic mesoderm morphogenesis will have an additive contribution on neural plate morphogenesis. To examine this possibility, we performed targeted injections using sub-optimal (half the amount of morpholino inducing a phenotype) amounts of Vangl2, Shroom3 and Wee2 morpholino. Specifically, we initially targeted the paraxial mesoderm with Wee2 morpholino, by injecting laterally of the dorsal marginal zone of 4 cell stage embryos. Subsequently half of those embryos at the 8-cell stage were injected with Vangl2 or Shroom3 morpholino at the 2 dorsal animal blastomeres at the. Uninjected 8 cell stage embryos were injected at the two dorsal blastomeres with suboptimal amount of only Shroom3 or only Vangl2 morpholino to obtain Vangl2 and Shroom3 single morphant embryos. Subsequently we assessed neural tube closure progression in control embryos, Wee2 single morphants, and Wee2/Vangl2 or Shroom3 double morphant embryos by quantifying neural plate width. This analysis revealed mild neural tube defects in Wee2, Vangl2 and Shroom3 single morphant embryos (Figures 7A–D). In contrast, in Wee2/Vangl2 and Wee2/Shroom3 double morphant embryos neural tube closure was significantly affected (Figures 7A–D). These data suggest that severe neural tube defects might emerge when a combination of mild defects in neural tube closure and somitic mesoderm morphogenesis co-exist. This might be the case, in severe neural tube defects emerging in otherwise normal LP/+ heterozygote mouse embryos when treated with chlorate (Nychyk et al., 2022). Chlorate is a potent inhibitor of CAG sulfation. Inhibition of CAG sulfation in LP/+ heterozygote mouse embryos, in which PCP signaling is partially lost, results in neural tube defects, which have been attributed to abnormal somite morphology.