Basal constriction during midbrain–hindbrain boundary morphogenesis is mediated by Wnt5b and focal adhesion kinase

Basal constriction is a cell shape change associated with zebrafish midbrain–hindbrain boundary constriction (MHBC) (Gutzman et al., 2008). This process contrasts with the widely studied mechanism of apical constriction (Martin and Goldstein, 2014). Following our initial identification of the process, basal constriction has been described in several other systems and developmental processes. It is associated with zebrafish and medaka optic cup morphogenesis (Bogdanović et al., 2012; Martinez-Morales et al., 2009; Nicolás-Pérez et al., 2016; Sidhaye and Norden, 2017), notochord cell elongation in Ciona, (Dong et al., 2011) and egg chamber elongation in Drosophila (He et al., 2010). Together these findings suggest that basal constriction is a conserved and fundamental morphogenetic process.

We previously demonstrated that basal constriction in the MHBC cells of the zebrafish neuroepithelium requires an intact basement membrane and is laminin-dependent (Gutzman et al., 2008). During optic cup morphogenesis, basal constriction has been demonstrated to require actomyosin contraction and is also dependent on laminin (Nicolás-Pérez et al., 2016; Sidhaye and Norden, 2017). However, the upstream signaling pathways that promote basal constriction have not been identified.

Since basal constriction at the zebrafish MHBC occurs within a small group of cells, one hypothesis is that a localized signaling process is involved. Wnt-PCP signaling is one candidate regulatory pathway, since this is crucial for many morphogenetic events, including gastrulation, convergent extension, cell migration, and cell adhesion (Ciani and Salinas, 2005) and has been studied during the development of the midbrain–hindbrain boundary (Buckles et al., 2004; Gibbs et al., 2017). Wnt5b is a PCP ligand and regulator of cell shape and movement. It is required during gastrulation (Jopling and den Hertog, 2005; Kilian et al., 2003; Lin et al., 2010), mesenchymal cell migration and adhesion (Bradley and Drissi, 2011), Xenopus bottle cell apical constriction (Choi and Sokol, 2009) and tail morphogenesis (Marlow et al., 2004). In this communication, we demonstrate expression of wnt5b at the zebrafish MHBC and find a connection between Wnt5b, Gsk3β and focal adhesion kinase (Fak), providing the first delineation of a signaling pathway required for basal constriction.

To delineate steps in basal constriction, we examined cell shape during midbrain–hindbrain boundary constriction (MHBC) by injecting wild-type embryos with membrane targeted GFP (mGFP) and imaging live embryos using confocal microscopy (Fig. 1A–D). MHBC morphogenesis takes place beginning at approximately the 18 somite stage (ss) and extending to the prim-6 stage. At the start, the neural tube is composed of a pseudostratified epithelium with established apical-basal polarity (Fig. 1A). We identified three sequential morphogenetic changes during MHBC formation. First, by 21 hpf MHBC cells become approximately 25% shorter than surrounding cells. Second, at 24 hpf, 3–4 cells within a single imaging plane at the point of deepest indentation of the MHBC each show basal constriction. Third, by 24 hpf MHBC cells expand apically by 60% relative to that of surrounding cells (Fig. 1A–H) (Gutzman et al., 2008, 2015). Three dimensional (3D) reconstruction of MHBC cells revealed that as the cells basally constrict and apically expand they become wedge-shaped (Fig. 1C–D,G–H). The average basal anteroposterior width of the MHBC cells decreases from 2.1 µm to less than 0.5 µm between 14 ss and prim-6 (Fig. 1I). The progression of MHBC cell shape change is summarized in Fig. 1J–M.

The ability to temporally separate morphogenetic steps supports the hypothesis that discrete molecular and mechanical processes underlie each step in MHBC formation. Our additional data are consistent with this hypothesis. For example, MHBC cell basal constriction can occur in the absence of apical expansion. Thus, in snakehead mutants that have reduced function in the atp1a1 gene encoding a Na⁺K⁺ ATPase, MHBC cells constrict basally, but fail to expand apically (Gutzman et al., 2008). In sleepy mutants, mapping to the lamc1 gene encoding for the gamma chain of Laminin 1-1-1, cell shortening occurs in the MHBC region, but both basal constriction and apical expansion fail to occur, indicating that cell shortening is not sufficient to drive subsequent steps and that apical expansion may depend on prior basal constriction (Gutzman et al., 2008). Additional studies support that independent molecular mechanisms govern various parts of MHBC morphogenesis. Prior to 24 ss, non-muscle myosin IIA and calcium signals are required to specifically shorten MHBC cells, but these molecules do not appear to have a direct role in mediating cell width (Gutzman et al., 2015; Sahu et al., 2017). In turn, non-muscle myosin IIB is not required for cell shortening, but is required to decrease cell width to fold the MHB (Gutzman et al., 2015; Sahu et al., 2017). Together, these data suggest that each step of MHB morphogenesis is regulated by separable molecular mechanisms.

We hypothesized that genes required for basal constriction would be expressed prior to the start of MHBC formation and that expression would be restricted to cells undergoing basal constriction or adjacent cells. In assessing the literature, we identified wnt5b expression as correlating with MHBC morphogenesis both temporally and spatially (Montero-Balaguer et al., 2006; Thisse and Thisse, 2005). Consistently, we demonstrated that wnt5b RNA was enriched at the MHBC using in situ hybridization (Fig. 2A–D). There is a low level of wnt5b expression throughout the embryo; however, expression increases at the MHBC shortly before morphogenesis begins and persists in this region throughout basal constriction (Fig. 2A–D). To determine the functional significance of Wnt5b in MHBC basal constriction, we used the established wnt5b antisense-morpholino modified oligonucleotide (MO) to inhibit function (Lele et al., 2001; Robu et al., 2007; Young et al., 2014). One-cell stage embryos were co-injected with control MO or wnt5b splice-site targeting MO with mGFP and live confocal imaging was employed to examine cell shape. Knockdown of wnt5b was associated with failure of basal constriction at the MHBC (Fig. 2E–F′). MHBC defects with Wnt5b knockdown could be a result of anomalous patterning in the MHB, which occurs earlier during development. However, expression of the patterning genes fgf8 and pax2, that are required for MHB formation, was unchanged in wnt5b morphants relative to control (Fig. S1). We also confirmed that wnt5b knockdown did not disrupt neuroepithelial cell apical-basal polarity (Fig. S1). These data suggest that requirement for wnt5b in basal constriction is not due to loss of early tissue patterning or polarity but is a later effect, perhaps directly impacting morphogenesis. At the tissue level, a slight constriction of the MHB was observable at prim-6 stage even after Wnt5b knockdown (Fig. 2F), suggesting additional forces or signals are involved in MHB morphogenesis. One possibility is that fluid secretion during embryonic brain ventricle inflation plays a role in promoting folding of the MHB tissue (Gutzman et al., 2008). The basement membrane, and specifically laminin, localized to the basal side of the cells, may also provide mechanical signals to influence tissue shape, which could be Wnt5b independent (Gutzman et al., 2008).

Wnt5b is a ligand that can activate non-canonical Wnt signaling through Rho and JNK, and can also act through inactivation of Gsk3β (De Rienzo et al., 2011; Niehrs and Acebron, 2010; Terrand et al., 2009; Torii et al., 2008). In zebrafish, Wnt5b functions as a negative regulator of Wnt/β-catenin activity (Westfall et al., 2003) and studies in Hydra suggest that, during evagination, Wnt5b may also be involved with crosstalk between the canonical and non-canonical Wnt signaling pathways (Philipp et al., 2009). We asked whether inhibition of Gsk3β is required for basal constriction, using a kinase-dead Gsk3β (dnGsk3β) that is an established dominant negative construct (De Rienzo et al., 2011; Yost et al., 1996). Supporting a connection between Wnt5b and inhibition of Gsk3β, co-injection of dnGsk3β together with the wnt5b MO prevented deficits in basal constriction seen after injection of wnt5b MO alone (Fig. 2G–H). Consistent with this, abnormalities in the gross morphology of the MHBC after injection of wnt5b MOs was prevented by expression of the dnGSK3β (Fig. S2). These data are consistent with a pathway in which Wnt5b regulates basal constriction through inhibition of Gsk3β.

Basal constriction at the MHBC and in the optic cup both require laminin (Gutzman et al., 2008; Nicolás-Pérez et al., 2016), a component of the underlying basement membrane, which interacts with integrins to regulate epithelial cell adhesion, migration and differentiation (Yamada and Sekiguchi, 2015; Yurchenco, 2015). Fak, a non-receptor tyrosine kinase, is a regulator of adhesion and cell migration that is activated through intracellular interactions with integrins (Parsons et al., 1994; Schaller, 2010). We therefore hypothesized that Fak plays a role in MHBC basal constriction. Since a primary mechanism for Fak activation is via autophosphorylation at Tyr397 (Schaller, 2010), we specifically hypothesized that autophosphorylated Fak^Y397 would be localized to the MBHC. An antibody specific to Fak autophosphorylation site Y397 stained both the apical and basal surfaces in the neural tube at 18 ss, 24 ss, and prim-6 (Fig. 3A–D).

We tested a role for Fak in basal constriction using knockdown with antisense-morpholino modified oligonucleotide injection. One-cell stage embryos were injected with control MO or a splice-site morpholino targeting the ptk2.1 gene encoding one of the two fak genes in zebrafish. fak MO efficacy was confirmed by RT-PCR and western blot analysis (Fig. S3A–C), and specificity was confirmed by rescue of fak MO injected embryos with co-injection of human FAK mRNA (Fig. S3). At the concentration used here, fak MO injections did not disrupt MHB tissue patterning or apical polarity markers (Fig. S1). fak morphants demonstrated disruption in MHB formation and abnormal basal constriction at the MHBC (Fig. 3E–F′). We tested activity of Fak^Y397 in regulation of basal constriction using a phosphomimetic mutation of Tyr397 to Glu397 (Fak^Y397E). Consistent with activation of Fak during basal constriction, co-injection of Fak^Y397E with fak MO was able to prevent abnormal basal constriction (Fig. 3G,G′).

We further tested the spatial and temporal requirement of Fak in the MHB to mediate basal constriction using an injection of a photoactivatable cyclic fak MO. With UV activation, the cyclic fak MO becomes linear and binds to its target site (Yamazoe et al., 2012). We injected wild-type embryos with cyclic fak MO, Kaede mRNA and mGFP mRNA. UV activation was performed at 14–16 ss, just before MHBC morphogenesis begins, in the MHB region as delineated by the change of Kaede from green to red. Basal constriction was disrupted after photoactivation of the cyclic fak MO, with no effect without activation or after UV treatment of the control MO-injected animals. These data show that Fak is required in the MHB region to mediate basal constriction (Fig. 3G–K′). We further determined that Fak functions cell autonomously using MHB targeted cell transplantation (Fig. S4). Together these data indicate that Fak activity is required for basal constriction, and that Fak functions in the cells of the region that is undergoing basal constriction, beginning just prior to the start of the process.

Since both Wnt5b and Fak are required for basal constriction, we asked whether Wnt5b functions upstream of Fak. To address this, we tested whether human FAK mRNA encoding the activated FAK^Y397E was able to prevent the basal constriction defect seen after Wnt5b inhibition. Indeed, we found that this mRNA was able to rescue basal constriction in wnt5b morphants (Fig. 4A–G). This effect was not general, as FAK^Y397E did not rescue basal constriction defects found in laminin mutants [(Gutzman et al., 2008) and Fig. S5], which is consistent with a structural role for laminin in the constriction process. These data indicate that Fak acts downstream of Wnt5b in activation of basal constriction at the MHBC.

Together, our results uncover key signaling factors contributing to basal constriction during MHBC morphogenesis. Our data point to a model in which Wnt5b signals locally at the MHBC as an early step in basal constriction, and acts together with more widespread Fak activation (Fig. 4H). We do not know whether or not Wnt5b and Fak function at the same time during this process, nor whether their activity is also necessary for the earlier cell shortening or the subsequent apical expansion. Future experiments will uncover the molecular details of this signaling interaction and the role in other steps of MHBC formation.

Zebrafish lines were maintained and embryo stages were determined as previously described (Kimmel et al., 1995; Westerfield, 2000). Zebrafish strains used include wild-type AB and sly^m86 (Schier et al., 1996). Due to the stages of development analyzed, we are unable to determine the sexes of the embryos. However, a large number of embryos were utilized for these experiments and each sex should be represented. All experimental procedures on live animals and embryos were reviewed and approved by the Institutional Animal Care and Use Committee of the Massachusetts Institute of Technology and were carried out in accordance with the recommendations in the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals.

All mRNA was in vitro transcribed with the mMessage mMachine kit (Thermo Fisher Scientific). Membrane-bound GFP (mGFP) mRNA was injected at 100–200 pg/embryo (kindly provided by J. B. Green, Dana-Farber Cancer Institute Boston, USA). Membrane-bound Cherry (mCherry) mRNA was injected at 50 pg/embryo (kindly provided by Dr Roger Tsien, University of California San Diego, USA). Photoconvertible Kaede mRNA was injected at 100 pg/embryo. pCS2+Kaede was kindly provided by Atsushi Miyawaki (RIKEN) (Ando et al., 2002). Human focal adhesion kinase (FAK) (accession number BC035404) was purchased from Open Biosystems (EHS1001-5481173) and constructs were generated by subcloning into the pCS2+ expression plasmid. mRNA was in vitro transcribed and injected at 200–250 pg/embryo. pCS2+FAK was used as the backbone to generate the FAK^Y397E phosphomimetic using QuickChange Site-Directed Mutagenesis (Agilent). For each mRNA injection and rescue experiment, all embryos were injected with equal amounts of total mRNA. This included total mGFP when needed for imaging by scanning confocal microscopy. All microinjection experiments were performed at least three times.

Live imaging of whole embryos was conducted using brightfield and fluorescent microscopy (SteREO Disvovery.V8, Zeiss). Live scanning confocal imaging was conducted as previously described (Graeden and Sive, 2009). Briefly, embryos were mounted inverted in 0.7% agarose and imaged using a 40× water immersion lens. Imaging was conducted using a Zeiss LSM510 or LSM720 scanning confocal microscope. Data were analyzed using Photoshop (Adobe) and Illustrator (Adobe) for cell outlines. 3D cell reconstruction was performed using 3D Doctor (Able Software). Individual cells at the MHBC were manually outlined in each z-section and rendered in 3D. A minimum of six embryos were imaged by scanning confocal microscopy and analyzed for basal constriction for each condition. Quantification of cell width was conducted using Imaris (Bitplane). The width of six cells at the MHBC from each of three embryos was measured at 300× zoom. Measurements were averaged and error bars reflect standard deviation for each condition.

Splice site-blocking morpholino (MO) antisense oligonucleotides were injected into embryos at the one-cell stage. Morpholinos and concentrations used are as follows: 3 ng/embryo of wnt5b MO targeting the exon5/6 splice donor 5′-TGTTTATTTCCTCACCATTCTTCCG-3′ (Kim et al., 2005; Robu et al., 2007); 0.75 ng/embryo of fak MO (ptk2.1) targeting the exon 12/13 splice donor 5′-GTGTGTTTGGGTTCTCACCTTTCTG-3′; non-specific sequence standard control MO 5′-CCTCTTACCTCAGTTACAATTTATA-3′ at the concentration equal to the test condition; and p53 MO 5′-GCGCCATTGCTTTGCAAGAATTG-3′ was co-injected at a concentration equal to 1.5 times the concentration of the test condition. Morpholinos were purchased from Gene Tools, LLC. Embryo phenotypes were classified as normal, mild, or severe. Normal embryos appeared as wild type. Mild embryos exhibited a consistent tail morphology defect, but basal constriction occurred normally. Severe phenotype embryos exhibited extensive tail defects and a basal constriction defect. Severe phenotype embryos were used for confocal imaging.

For photoactivatable morpholino experiments, we injected one-cell stage embryos with 1 ng/embryo of splice site-targeting cyclic fak MO (ptk2.1) 5′-GTGGGTGCTAACTGTCCGTCATATT-3′. The fak MO was cyclized with a photocleavable linker as previously described (Yamazoe et al., 2012) and remains inactive until ‘uncaging’ by UV light. Linker photolysis reverts the MO to a linear oligonucleotide that can target the fak splice site. Embryos were co-injected with mGFP mRNA and Kaede mRNA together with either cyclic fak MO or control MO at the one-cell stage. Region and time specific UV activation was conducted at the 10–16 somite stage on cells located in the prospective MHB using a Zeiss Axioplan2 fluorescent microscope with a UV filter and adjustable iris. The tissue region that was activated by UV light is visible with the Kaede color change from green to red. Only cells that were photoconverted at the MHBC were analyzed for basal constriction as described.

Antisense and sense RNA probes containing digoxigenin-11-UTP were synthesized from linearized plasmid DNA for wnt5b was obtained from Addgene #21282 (Stoick-Cooper et al., 2007). Standard methods for hybridization and for single color labeling were used as described (Sagerstrom et al., 1996). After staining, embryos were de-yolked, flat-mounted in glycerol and imaged with a Nikon compound microscope or a Zeiss Discovery V8.

Embryos were fixed in 4% paraformaldehyde; blocked in 2% normal goat serum, 1% BSA, and 0.1% Triton-X100 in PBT; incubated overnight at 4°C in primary antibody (anti-phosphoY397-FAK, 44-624 BioSource, Life Technologies), 1:200; then incubated in secondary antibody (goat anti-rabbit IgG conjugated with Alexa Fluor 488, Invitrogen, 1:500). Embryos were de-yolked and mounted in glycerol. Images were collected using scanning confocal microscopy (Zeiss LSM510 or 710) and analyzed using Photoshop (Adobe).

We thank our Sive lab colleagues for useful criticism and Oliver Paugois for excellent fish care. We would like to acknowledge the W. M. Keck Foundation Biological Imaging Facility at the Whitehead Institute for advice and imaging assistance. pCS2+Kaede was kindly provided by Atsushi Miyawaki (RIKEN).