Neural tube defects by NUAK1 and NUAK2 double mutation


  • Tomomi Ohmura,

    1. Laboratory for Vertebrate Body Plan, Center for Developmental Biology, RIKEN Kobe, Kobe, Japan
    2. Department of Physiology and Cell Biology, Kobe University Graduate School of Medicine, Hyogo, Japan
    Search for more papers by this author
  • Go Shioi,

    1. Laboratory for Animal Resources and Genetic Engineering, Center for Developmental Biology, RIKEN Kobe, Kobe, Japan
    Search for more papers by this author
  • Mariko Hirano,

    1. Laboratory for Vertebrate Body Plan, Center for Developmental Biology, RIKEN Kobe, Kobe, Japan
    Search for more papers by this author
  • Shinichi Aizawa

    Corresponding author
    1. Laboratory for Vertebrate Body Plan, Center for Developmental Biology, RIKEN Kobe, Kobe, Japan
    2. Laboratory for Animal Resources and Genetic Engineering, Center for Developmental Biology, RIKEN Kobe, Kobe, Japan
    • Laboratory for Vertebrate Body Plan, Center for Developmental Biology, RIKEN Kobe, 2-2-1 Minatojima-minami, Chuo-ku, Kobe, Hyogo 651-0055, Japan
    Search for more papers by this author


Background: NUAK1 and NUAK2, members of the AMP-activated protein kinase family of serine/threonine kinases, are prominently expressed in neuroectoderm, but their functions in neurulation have not been elucidated. Results: NUAK1 and NUAK2 double mutants exhibited exencephaly, facial clefting, and spina bifida. Median hinge point was formed, but dorsolateral hinge point formation was not apparent in cranial neural plate; neither apical constriction nor apico-basal elongation took place efficiently in the double mutants during the 5–10-somite stages. Concomitantly, the apical concentration of phosphorylated myosin light chain 2, F-actin, and cortactin was insignificant, and development of acetylated α-tubulin-positive microtubules was poor. However, the distribution of F-actin, cortactin, Shroom3, Rho, myosin heavy chain IIB, phosphorylated myosin light chain 2, α-tubulin, γ-tubulin, or acetylated α-tubulin was apparently normal in the double mutant neuroepithelia at the 5-somite stage. Conclusions: NUAK1 and NUAK2 complementarily function in the apical constriction and apico-basal elongation that associate with the dorsolateral hinge point formation in cephalic neural plate during the 5- to 10-somite stages. In the double mutant neural plate, phosphorylated myosin light chain 2, F-actin, and cortactin did not concentrate efficiently in apical surfaces, and acetylated α-tubulin-positive microtubules did not develop significantly. Developmental Dynamics 241:1350–1364, 2012. © 2012 Wiley Periodicals, Inc.


NUAK/Omphk are members of the AMPK (AMP-activated protein kinase) family of serine/threonine kinases (Manning et al., 2002). Vertebrates have two paralogues: NUAK1/Omphk1/ARK5 and NUAK2/Omphk2/SNARK (Hirano et al., 2006). AMPKs are activated by a tumor suppressor LKB1 (Lizcano et al., 2004), but NUAK1 is also activated by Akt (Suzuki et al., 2003b) and NDR2 (Suzuki et al., 2006). NUAK1 is known as the tumor cell survival factor; it is activated in colorectal tumors and promotes invasiveness (Kusakai et al., 2004; Suzuki et al., 2004a). NUAK1 phosphorylates ATM, which activates p53 (Suzuki et al., 2003b); it induces membrane type 1 matrix metalloprotease (MT1-MMP) expression by rapamycin-sensitive signaling, and promotes cell detachment or suppresses cell adhesion (Suzuki et al., 2004b). NUAK2 also regulates cell–cell and cell–matrix adhesion (Suzuki et al., 2003c). Furthermore, NUAK1 and NUAK2 are reported to suppress apoptosis (Suzuki et al., 2003a; Legembre et al., 2004).

Recently, Zagorska et al. (2010) reported that NUAK1 phosphorylates and suppresses myosin phosphatase targeting-1 (MYPT1); the phosphorylation induces MYPT1 binding to 14-3-3, thereby inhibiting the MYPT1 interaction with myosin and leading to increased phosphorylation of myosin light chain 2 (MLC2) and activation of myosin II. Actomyosin cytoskeleton provides contractile forces to both assemble and disassemble cell–cell and cell–matrix contacts. Phosphorylation of MLC2 plays essential roles in the regulation of cell polarity, adhesion, migration, tissue architecture, and other contractile functions in non-muscle cells (Vicente-Manzanares et al., 2009). NUAK2 is also reported to increase MLC2 phosphorylation by associating with MRIP (myosin phosphatase Rho-interacting protein) and promotes formation of stress fibers (Vallenius et al., 2011). However, NUAK1 and NUAK2 functions in vivo have not been well demonstrated. Previously, we reported that mouse NUAK1 mutants exhibit ventral body wall defects. NUAK1 and NUAK2 are most prominently expressed in the entire neuroectoderm, but their functions in neurulation have not been elucidated. In this report, we examined neurulation defects in NUAK1 and NUAK2 (NUAK1/2) double mutants.

Neurulation starts with neural plate shaping. The neural plate is initially elliptical, but is converted to an elongated keyhole-shaped structure with broad cranial and narrow spinal regions (Copp et al., 2003). The anterior-posterior elongation of the plate is followed by bending to form a tube; the lateral edges or neural folds rise, converge at the dorsal midline, and fuse to form the neural tube. The bending occurs at two principal sites, the median hinge point (MHP) and the paired dorsolateral hinge points (DLHPs). The upper spine bends only at MHP, the lower spine at DLHPs, while the intermediate spine and cranial region at both MHP and DLHPs. Rostro-caudally, the neural tube closure is initiated at three points in mice. The closure at the hindbrain/cervical boundary takes place at the 6- to 7-somite stage (E8.5) and spreads rostrally and caudally from this site (closure 1). A second de novo closure event (closure 2) occurs at the forebrain/midbrain boundary, also spreading rostrally and caudally. Closure also initiates separately at the rostral extremity of the forebrain (closure 3), spreading caudally to meet the rostral spread of fusion from closure 2.

The main driving force for the neural plate elongation is convergent extension, which is regulated by the non-canonical Wnt/planar cell-polarity (PCP) pathway (Wallingford and Harland, 2002). Mutants of genes implicated in the PCP pathway exhibit neural tube defects (NTDs), especially craniorachischisis (CRS), in which closure 1 fails (Kibar et al., 2001; Hamblet et al., 2002; Curtin et al., 2003; Montcouquiol et al., 2003; Harris and Juriloff, 2007, 2010). Columnar neuroepithelial cells adhere tightly to each other through the junctional complexes such as adherens junctions (AJs) located close to the apical surface. The structures are linked to the actin cytoskeleton to form a dense, submembranous actin belt that serves to link adjacent cells and regulate cell shape, migration, and polarity (Gumbiner, 1996). Several lines of knockout mice lacking cytoskeleton-associated proteins, such as p190 RhoGAP, Shroom3, Mena, MARCKS, and vinculin, have shown NTDs (Wu et al., 1996; Xu et al., 1998; Hildebrand and Soriano, 1999; Lanier et al., 1999; Brouns et al., 2000). The actin cytoskeletons are often jointly regulated with microtubule cytoskeletons (Rodriguez et al., 2003), and the roles of microtubule networks have also been suggested in neural tube closure (Kinoshita et al., 2008; Suzuki et al., 2010). However, molecular mechanisms of cytoskeletal reorganization, directed by the PCP and other pathways, during neurulation remain issues for further exploration, and the molecular mechanisms underlying MHP and DLHP formation are poorly understood.

In this study, we report that the NUAK1/2 double mutants exhibit neural tube defects. MHP is formed, but DLHP formation is not apparent at the cranial level; neither apical constriction nor apico-basal elongation occurs efficiently in the double mutant cranial neural plate. The elucidation of the NUAK1/2 functions should expand our understanding of the molecular mechanisms of cytoskeletal reorganization in neuroepithelial cells during neurulation.


NUAK2 Targeting and Expression

Gene targeting was designed to disrupt the NUAK2 gene by replacing the coding domain of the first exon with a lacZ-neo cassette or lacZ gene (Fig. 1A), as our previous targeting of the NUAK1 gene (Hirano et al., 2006); β-galactosidase (βGal) should be expressed in place of NUAK2. No change in βGal expression was found between NUAK2 mutants, which have neo cassette (lacZ-neo allele) and in which the Neo cassette was deleted by flippase (lacZ allele). The details of βGal expression were examined with the lacZ mutants, but most of the phenotype analysis was conducted with the lacZ-neo mutants. RT-PCR analysis of the NUAK2 mRNA expression with primers at exon 2 and exon 5, and Western blot analysis with an antibody against the C-terminal part of NUAK2 indicated the null nature of the mutation (Fig. 1C), as true of the NUAK1 mutation (Hirano et al., 2006).

Figure 1.

NUAK2 targeting strategy. A: Schematic representation of NUAK2 wild type (wt) allele, targeting vector, lacZ-neo and lacZ targeted allele. Box in wt allele, the first coding exon; open box, untranslated 5′ region; filled box, coding region. Gray triangles, loxP sequences; small black triangles, frt sequences. LacZpA is lacZ gene with rabbit β globin polyadenylation signal; Neo, neomycin resistant gene directed by PGK1 promoter; pA, PGK1 polyadenylation signal; DT-A, diphtheria toxin A-fragment gene directed by MC1 promoter. 3′ probe indicates the location of the probe used for Southern blot analysis in (B, left); P1–P3, primers for routine genotyping of the allele shown in B (right). B: Genotyping of the allele; an example of Southern blot analysis (left) and routine genotyping by PCR (right) with tail DNAs. C: RT-PCR and Western blot analyses of NUAK2 expression in E8.5 embryos, indicating the null nature of the mutation. The primers used detect the exons 2–5 region, and the antibody used recognizes the C-terminal region (523–639 aa). P and N, positive and negative controls.

The βGal expression was not detected in E 6.5 embryos at the onset of gastrulation (see Supp. Fig. S1a, which is available online), but occurred in anterior neuroectoderm induced at E 7.5 (Fig. 2). In E 8.5 and E 9.5 embryos, the βGal expression from NUAK2 locus was also exclusively found in the entire neuroectoderm. This βGal expression was exactly the same as the NUAK2 expression by whole mount RNA in situ hybridization (Hirano et al., 2006). NUAK2 expression contrasts with NUAK1 expression, which is also present in a variety of non-neural tissues including the ventral body wall (Fig. 2) (Hirano et al., 2006).

Figure 2.

NUAK1 and NUAK2 expression as represented by the expression of βGal knocked in each locus. a, c, f, i, k, n: Lateral whole mount views. b. d, g, j, l, o: Parasagittal sections. e, h, m, p: Frontal sections at the planes indicated by dotted lines in c, f, k, n.

NUAK1/2 Double Mutant Phenotype

NUAK2 single mutants exhibited exencephaly at the frequency of about 40% (Fig. 3Ab, e, h); NUAK1 single mutants do not have NTDs as reported previously (Hirano et al., 2006). NUAK1/2 double mutants had facial clefting and spina bifida as well as exencephaly at 100% frequency among more than 100 embryos examined (Fig. 3Ac). Neither closure 2 nor 3 took place, while closure 1 at the hindbrain/cervical boundary took place and spread caudally but not rostrally (Fig. 3Af, i). Additionally, the spinal cord that achieved closure failed to establish the correct straight tubular shape and was undulated (Fig. 3Ba,b). The double mutants also exhibited eye defects at 100% frequency; neural retina and retina pigment epithelium were overgrown with the failure of optic fissure closure (Fig. 3Bc, d). These mutants also had body wall defects seen in NUAK1 single mutants (Fig. 3Bg, h) (Hirano et al., 2006). However, somites and notochord were histologically normal (Fig. 3Ba, b, e–h); Cerl-positive nascent somites and Uncx4.1-positive somites normally developed in the NUAK1/2 double mutants (Fig. 3Ca–d). Branchial arches, heart, and other tissues were also normally formed in the double mutants (Fig. 3Be–h). NUAK1/2 double mutant embryos were viable at all embryonic stages and survived to term. NTDs can be mistakenly described in embryos that are developmentally retarded prior to mid-gestation embryonic lethality (Copp et al., 2003), but the NUAK1/2 double mutant NTDs are apparently not secondary to mid-gestation lethality.

Figure 3.

NUAK1/2 double mutant phenotype. A: Whole mount lateral (a–c), frontal (d–f), and dorsal (g–i) views of E12.5 (a–c) and E9.5 (d–i) embryos of each genotype. (+/+, +/+) indicates wild type (NUAK1+/+NUAK2+/+), (+/ +, −/−) NUAK2 single mutant (NUAK1+/+NUAK2lacZ-neo/lacZ-neo) and (−/−, −/−) NUAK1/2 double mutant (NUAK1lacZ-neo/lacZ-neoNUAK2lacZ-neo/lacZ-neo) embryos. B: Histological views of wild type and NUAK1/2 double mutants. a, b: Horizontal sections of the spinal cord closed at E10.5. c, d: Eye defects at E12.5. e–h: Frontal sections at thoracic (e, f) and abdominal (g, h) levels at E12.5. C: Development of somitic mesoderm. Whole mount views of (a, b) Cerl-positive nascent somites at E9.5 and (c, d) Uncx4.1-positive somites at E8.5. NUAK1 single mutants do not have NTDs (Hirano et al., 2006), but NUAK2 single mutants exhibited exencephaly at the frequency of about 40%. NUAK1/2 double mutants had facial clefting and spina bifida as well as exencephaly at 100% frequency. NUAK1 single mutants exhibit a body wall defect (Hirano et al., 2006), and thus the NUAK1/2 double mutants had this defect (Hirano et al., 2006). The double mutants also exhibited eye defects. However, other tissues were apparently normal, and NUAK1/2 double mutant embryos were viable at all embryonic stages and survived to term.

NUAK has been reported to regulate cell behaviors such as cell proliferation, cell differentiation, and apoptosis (Legembre et al., 2004; Suzuki et al., 2004a; Hou et al., 2011; Namiki et al., 2011). In mouse, the neuroepithelium is entirely proliferative during neurulation; neuronal differentiation occurs only after neural tube closure. Neural plate closure has been known to be regulated by proper cell proliferation. Both decrease and increase in cell proliferation and premature differentiation into neurons could cause NTDs (Copp et al., 2003). Cranial neurulation is also sensitive to the alternation in the extent of neuroepithelial cell death (Copp et al., 2003). Therefore, the cell behaviors were examined in neuroepithelia at the midbrain level. However, no changes in PH3-positive cells were apparent in the NUAK1/2 double mutant neuroepithelia at the 5- and 11-somite stages (Fig. 4a, b). Counting more than 1,000 cells of seven sections from seven embryos, the average number of PH3-positive cells was 8.6 ± 2.7 and 8.9 ± 3.3 cells per 100 neuroepithelial cells in wild type at the 5–6- and 9–11-somite stages, respectively. It was 8.2 ± 2.7 and 8.4 ± 2.2 in the double mutant. Furthermore, no premature differentiation of MAP2-positive neurons was observed in the double mutants (Fig. 4c–e). No increase in caspase3-positive cells was also found in the double mutants (Fig. 4f, g, see legend).

Figure 4.

Neuroepithelial cell behaviors in NUAK1/2 double mutants. a, b: PH3 staining for cell proliferation. c, d: MAP2 staining for neural differentiation. e: A positive control of MAP2 staining. f, g: Caspase3 staining for cell death. (h, i) N-cadherin and (j, k) β1 integrin staining for cell adhesions, and (l, m) laminin staining for basal lamina at stages indicated in parenthesis (s, somite stage), respectively. Previous studies in culture expected, but NUAK1/2 double mutant neuroepithelial cells did not show, apparent decrease in cell proliferation, premature neuronal differentiation, or increase in cell death. Arrowheads in f, g indicate cell death prominent at the tips of the neural folds in both wild type and double mutant neural plate. In other parts of the midbrain neural plate, no cell death was apparent; the frequency of caspase3-positive cells was on the order of one in 500 cells (examined in six sections from 2 embryos) of either wild type or NUAK1/2 double mutants. There were also no apparent changes in cell–cell adhesions or cell–ECM adhesions. n: The planes of sections in the 5–6-, 10–11-, or 21–22-somite stage of wild type or double mutant midbrain are indicated by dotted lines, except for the planes prepared for caspase 3 staining shown by solid lines.

NUAK has been also reported to regulate the cell–cell and cell–matrix adhesions (Suzuki et al., 2003c, 2004b). Changes in cell–cell adhesions and loss of basal lamina components in neuroepithelial cells have been reported to cause NTDs (Miner et al., 1998; Arikawa-Hirasawa et al., 1999; Morita et al., 2010). However, there were no changes in N-cadherin or β1 integrin distribution found between wild type and double mutant neuroepithelial cells (Fig. 4h–k), and laminin- and collagen V-positive basal lamina were normally present in the double mutants (Fig. 4l, m; Supp. Fig. S1l–o).

Therefore, NUAK1/2 double mutant neuroepithelia did not demonstrate NUAK functions in cell proliferation, differentiation, cell death, cell-cell adhesions or cell-matrix adhesions, although previous studies in culture suggested them. NUAK1 and NUAK2 expressions are uniform in neural plate, and it is also less likely that a locally subtle change, for example in prospective DLHP, in either of these cell behaviors are at the root of the failure in the NTDs of NUAK1/2 double mutants.

MHP Is Formed But DLHPs Are Not

In the cranial region, rising neural folds initially adopt a biconvex morphology, and this process was apparently normal in the NUAK1/2 double mutants (Fig. 5A). In the bending process, histologically NUAK1/2 double mutant neural plate formed MHP, but DLHP formation was not apparent (Fig. 5A). In the lower spine where both MHP and DLHPs are formed in wild type, MHP was formed but DLHPs were also not apparent in NUAK1/2 double mutants (Fig. 5Ba–d). In the more caudal spine where DLHPs are formed but MHP is not in the wild type, the double mutant neural plate scarcely bent (Fig. 5Be–h). MHP is characterized by the basally wedge-shaped cells in which the nuclei occupies the basal position and by the decrease in cell proliferation (Smith and Schoenwolf, 1988). No significant change was apparent in the frequency of the wedge-shaped cells in the NUAK1/2 double mutant cephalic MHP (Fig. 5C). Furthermore, cell proliferation was low in NUAK1/2 double mutant MHP as in wild type MHP. Average frequency of PH3-positive cells was 3.8±2.8 cells and 4.0±2.8 cells per 100 cells in the wild type and double mutant MHP at the 9–11-somite stage, respectively, when 240 and 150 cells of eight and five sections from five embryos were counted, respectively (see above for the frequency in other parts of the neural plate). The MHP cells are also characterized by Shh-positive cells; they were normally present in the double mutant ventral midline (Fig. 6g, h). Taking these results together, we consider that MHP is formed normally in the NUAK1/2 double mutants, and NTDs in NUAK1/2 double mutants may be associated with the failure in DLHP formation. It has been reported that the intermediate spine closes in mutants such as Zic2 mutants that do not develop DLHP (Ybot-Gonzalez et al., 2007). The present studies focus on DLHP formation at the cranial region. The onset of the defects must precede the 10-somite stage.

Figure 5.

Hinge point formation in NUAK1/2 double mutants. A: Histological features of neurulation at midbrain level in NUAK1/2 double mutants at stages indicated; frontal sections. Histologically, defects were not marked at the 5–7-somite stage; MHP was apparently formed in NUAK1/2 mutants. However, at the 9-somite stage, the bending starts in wild type with the formation of DLHPs (arrowheads), but does not in NUAK1/2 double mutants. See Figure 4n for the planes of sections at each stage. B: Histological features of neurulation at lower spine levels; frontal sections. a–d: The features at the level where both MHP and DLHPs are formed, and (e–h) at the level where DLHPs are formed but MHP is not, in wild type embryos; the levels are indicated (i, j) by solid and dotted lines, respectively. C: The frequency of spindle-shaped (S), wedge-shaped (W), inverted wedge-shaped (IW), and globular (G) cells in 140 MHP cells at the 10–12-somite stages; each shape of the cells is as defined by Schoenwolf and Franks (1984). Twenty cells at the midline of the neural plate were counted on each section at the midbrain level. The frequency of each type of cell was also determined at the sites indicated by arrowheads and arrows (Ae, f) at the 10-somite stage. No significant differences in the frequency of either type of cell were apparent among these sites or between wild type and double mutant plates.

Figure 6.

Marker analyses for DLHP defects in NUAK1/2 double mutants. (a–d) twist and (e, f) Crabp1 expression in head mesenchyme. g, h: Shh expression in midline. i, j: Bmp2 expression in surface ectoderm adjacent to neural plate (arrowheads). k, l: Msx1 expression in the lateral edges of the neural plate, adjacent surface ectoderm, and head mesenchyme (arrowheads), at stages indicated in parenthesis, respectively. All are present normally in NUAK1/2 double mutants. a, b give whole mount views, and all others are sections at the midbrain level. See Figure 4n for the planes of sections at each stage.

In the cranial region, the neural tube closure has been suggested to depend on the cranial mesenchyme (Chen and Behringer, 1995; Zhao et al., 1996) or the cranial neural crest migration (Ewart et al., 1997). Histologically, the head mesenchyme was normally formed in NUAK1/2 double mutants (Fig. 5A). Twist, Cart1, and Crabp1-positive neural crest cells were normally migrating at the 10-somite stage when the neural tube defects became distinct (Fig. 6a–f; Supp. Fig. S1b–g). Together with the fact that NUAK2 is not expressed in tissues other than neuroepithelium, the NTDs are probably caused by a defect that acts autonomously within the cells comprising the neural plate.

At the spine level, Shh has been known to be inhibitory to the DLHP formation (Echelard et al., 1993; Ybot-Gonzalez et al., 2002). Ventralization of neural plate by Shh overexpression in axial mesoderm and floor plate of the neural plate prevents the differentiation of dorsal cells that are necessary for DLHP formation. It inhibits this formation and causes NTDs (Echelard et al., 1993; Ybot-Gonzalez et al., 2002; Copp et al., 2003). In PCP mutants such as Vangl2 and Dvl1/2 double mutants, Shh-positive notochord and floor plate are expanded (Gerrelli and Copp, 1997; Wang et al., 2006). Dorsal cell fates are not specified, as indicated by the lack of expression of genes such as Msx1 and Wnt3a, in mouse mutants of Opb and Zic2, products of which suppress Shh signaling (Eggenschwiler and Anderson, 2000; Nagai et al., 2000). DLHP formation at the spine level is also suggested to be affected by BMP2 expressed at the surface ectoderm adjacent to the neural folds and noggin expressed at the lateral edges or tips of the neural folds (Ybot-Gonzalez et al., 2007). Shh and BMP signaling may also operate in the neural tube closure at the cranial region (Hui and Joyner, 1993; Gunther et al., 1994; Nagai et al., 2000; Huang et al., 2002; Eom et al., 2011). However, in the cranial region of the NUAK1/2 double mutants, Shh expression in axial mesoderm and floor plate was normal (Fig. 6g, h). BMP2 and noggin expression was also normally found (Fig. 6i, j; Supp. Fig. S1h, i). At the same time, Msx1 and Wnt3a expression in dorsal neural plate (Nagai et al., 2000; Eggenschwiler et al., 2001) was also normally present in NUAK1/2 double mutants (Fig. 6k, l; Supp. Fig. S1j, k), suggesting normal dorso-ventral patterning in their neural plate.

Failure in Apical Constriction

During the neural plate bending, the neuroepithelial cells constrict apically, and their apical surface area is reduced. This constriction is linked to the apico-basal cell elongation; the neural plate thickens, increasing cell height and decreasing the plate width (Colas and Schoenwolf, 2001; Copp et al., 2003). F-actin and cortactin become prominent in the apical surfaces of the wild type neural plate with bending during the 5- to 10-somite stages; they concentrate in the apical surfaces with constriction (Fig. 7A). At the 5-somite stage, F-actin and cortactin distribution in the NUAK1/2 double mutants was comparable with that in the wild type neural plate, but their increase in apical surfaces at the 10-somite stage was not significant in the absence of neural plate bending. At the 5-somite stage, there was no significant difference in the apical size or apico-basal height of neural epithelial cells between wild type and NUAK1/2 double mutants (Fig. 7Ba,b, C). However, subsequently the NUAK1/2 double mutant neuroepithelial cells did not decrease their apical surface area efficiently or significantly thicken apico-basally (Fig. 7Bc–f, C). The average size of the apical surface of the wild type cells at the 13-somite stage was about 12% that at the 5-somite stage, while that of mutant cells at the 13-somite stage was about 56% that at the 5-somite stage (Fig. 7Ca). The wild type neutral plate thickened 2.0-fold while the double mutant thickened 1.4-fold by the 10-somite stage (Fig. 7Cb).

Figure 7.

Lack of apical constriction and cell thickening in NUAK1/2 double mutant neural plate. A: Frontal views. a, b, g, h: F-actin expression. c, d, i, j: Cortactin expression. e, f, k, l: Their merged views at stages indicated. At the 5-somite stage, they are normally present in the apical side of double mutant neural plate, though at the 10-somite stage their intensities did not increase significantly, concordant with the lack of neural plate bending in the double mutants. See Figure 4n for the planes of sections at each stage. B: Apical views. a–f: F-actin staining at stages indicated. The inset indicates the positions at examinations; the neural plate was split into two halves at the midline, and the F-actin staining was compared at a midbrain position indicated by a red circle. C: Quantification of (a) apical cell surface area and (b) apico-basal cell thickness at indicated stages. The apical surface area of each cell was determined on a total of 120 cells (40 cells in each section from three embryos) using ImageJ software. The thickness was determined by frontal sections perpendicular to the neural plate at the midbrain level.

Distribution of Actomyosin Network Components

Neuroepithelial cells are polarized apico-basally. Apical constriction of the neuroepithelial cells is considered to be caused by activation of the actomyosin network associated with cadherin complex at AJs apico-laterally located (Morita et al., 2010). Myosin regulatory light chain 2 (MLC2) of myosin II is a key molecule that monitors the assembly-disassembly balance and contractility of actin fibers. Rho has been implicated in the formation of AJs and in actomyosin contraction through phosphorylation of MLC2 (Jaffe and Hall, 2005). Shroom 3, an actin-binding protein, is localized at the AJs and recruits Rho kinase to the junctions to contract actin filaments (Hildebrand, 2005; Nishimura and Takeichi, 2008). NUAK1/2 double mutant neuroepithelial cells are polarized normally. ZO1, aPKC, and PALS1 were tightly localized at apical regions of neural plate cells at the 5-somite stage; their distribution was also unchanged at the 10-somite stage (Fig. 8A, Supp. Fig. S2). In addition, there were no marked differences in the distribution of Shroom3 or Rho between wild type and NUAK1/2 double mutant neuroepithelial cells at the 5-somite stage or in their apical accumulation at the 10-somite stage (Fig. 8Ba–d, Supp. Figs. S1, S2). Apically positioned myosin heavy chain IIB (MHCIIB) is markedly reduced in Shroom3 mutants in which neural plate bending does not take place (Hildebrand, 2005), but MHCIIB was normally found in apical regions of the double mutant neuroepithelia at both the 5- and 10-somite stages (Fig. 8Be, f; Supp. Fig. S1).

Figure 8.

Cytoskeletal systems in NUAK1/2 double mutant neuroepithelium. A: Cell polarization. a–d: Expression of ZO1. e–h: Expression of aPKC. i, j: Expression of PALS1 by immunohistology at stages indicated in parenthesis. No differences are apparent in expression between wild type and double mutant neuroepithelia. B: Actomyosin networks. Distribution of Shroom3 (a, b), Rho (c, d), MHCIIB (e, f), and pMLC2 (g–j) by immunohistology at stages indicated. No changes in Shroom3, Rho, or MHCIIB distribution are apparent at either the 5-somite stage or 10-somite stage. However, the increase in pMLC2 intensity in the apical side of the double mutant neuroepithelia is less apparent, concordant with the lack of apical constriction. k: Western blot analysis of MLC2 and pMLC2 amounts in E9.5 brain. There were no marked changes in the NUAK1/2 double mutants. C: Tubulin networks. Distribution of α-tubulin (a, b), γ-tubulin (c, d), and acetylated α-tubulin (e–j) by immunohistology at indicated stages. There were no changes in α-tubulin or γ-tubulin distribution. However, development of acetylated α-tubulin-positive stable microtubules is poor, concordant with the lack of neural plate thickening. Magnified views are given in Supp. Figure S2. See Figure 4n for the planes of sections at each stage.

Phosphorylated MLC2 (pMLC2) is dephosphorylated by the myosin light chain phosphatase (MLCP), which is suppressed by phosphorylation of MLCP by ROCK, NUAK1, and NUAK2 (Koga and Ikebe, 2008; Zagorska et al., 2010; Vallenius et al., 2011). NUAK1 is reported to enhance phosphorylation of MLC2 in human embryonic kidney (HEK) 293 cells (Zagorska et al., 2010) and NUAK2 in HeLa cells (Vallenius et al., 2011). At the 5-somite stage, pMLC2 is found sparsely both at the basal and apical aspects of wild type neuroepithelial cells. It becomes intense at the apical region by the 10-somite stage in a mesh-like pattern (Fig. 8Bg, i; Supp. Fig. S2). At the 5-somite stage, there was no significant difference in the pMLC2 distribution by the NUAK1/2 double mutation (Fig. 8Bg, h). At the 10-somite stage, pMLC2 was less intense in the apical region of the double mutant neuroepithelium that inefficiently underwent the apical constriction (Fig. 8Bi, j, Supp. Fig. S2). Western blotting with E9.5 brain suggested no apparent decrease in the total amount of either MLC2 or pMLC2 (Fig. 8Bk).

Distribution of Microtubule Components

The apico-basal cell elongation has been suggested to be mediated also by microtubules (Smith and Schoenwolf, 1997; Suzuki et al., 2010). Microtubules and actin cytoskeletons are often jointly regulated (Rodriguez et al., 2003), and Shroom3, which drives apical constriction, is known also to induce a redistribution of γ-tubulin, which is a key player in the assembly of the robust MT arrays observed in elongating neural epithelial cells (Lee et al., 2007). At the same time, changes in cell height could occur independently of and preceding apical constriction. The defects in apicobasal cell elongation might be the cause of the failure in subsequent apical constriction (Lee et al., 2007). However, no changes in α-tubulin or γ-tubulin distribution were apparent in the NUAK1/2 double mutants (Fig. 8Ca–d, Supp. Fig. S2). γ-tubulin exhibited centrosome-like foci at the apical surfaces normally in the double mutants. In wild type cranial neural plate, stable microtubules begin to be formed at the 5-somite stage as evidenced by acetylated α-tubulin staining. The microtubules started to be formed normally at the 5-somite stage in the double mutants (8Ce, f, Supp. Fig. S2). Subsequently along with the apico-basal elongation, the development of stable microtubules is marked in wild type cephalic neural plate (Fig.8Cg, i). However, the development of acetylated α-tubulin-positive microtubule was poor in the double mutants (Fig. 8Ch, j), which underwent inefficient neural plate thickening.

This study showed that NTDs in NUAK1/2 mutants are associated with the failure in DLHP formation. At the cranial level, it may be due to the inefficient apical constriction and apico-basal elongation of neuroepithelial cells. Both did take place but inefficiently. The defects appeared to precede the 10-somite stage histologically. During the 5- to 10-somite stages, normally F-actin, cortactin, and pMLC2 are concentrated in apical surfaces with apical constriction, but this did not take place in the double mutants. The development of acetylated α-tubulin-positive stable microtubules was poor in the NUAK1/2 double mutants at the 10-somite stage with the inefficient apico-basal elongation. However, all these were normally found at the 5-somite stage, and molecular mechanisms or NUAK1 and NUAK2 direct targets causing these defects remain to be determined by future studies.

The previous studies that indicated NUAK1 and NUAK2 enhance MLC2 phosphorylation in cultured cells anticipated that the loss of or the decrease in pMLC2 would be the most plausible cause of the NTDs in NUAK1/2 double mutants, but this study could not demonstrate this. The increase in the apical intensity of pMLC2 was poor in the double mutants at the 10-somite stage, but this may be the result rather than the cause of the inefficient apical constriction. The total pMLC2 amount was probably unchanged in the cranial neural plate. However, the roles of NUAK1 and NUAK2 in the actomyosin network for apical constriction, together with their roles in microtubule assembly, must be examined in more detail by future studies. MLC2 phosphorylation must be regulated by many factors other than NUAK1 and NUAK2 during neurulation, and even in their absence the phosoprylation of MLC2 might be largely unchanged (Zagorska et al., 2010). It might be the increase in the phosphorylation of a subset of MLC2 for which NUAK is responsible during neurulation. For example, NUAK might specifically phosphorylate MLC2 along the AJs distributed toward the dorsoventral direction of the neural plate bending (Nishimura and Takeichi, 2008). At the same time, there have been suggestions that actin cytoskeletons do not drive the wedging of cells at hinge points, but rather stabilize the shape of the neural folds or newly formed neural tube (Colas and Schoenwolf, 2001). Actin cytoskeletons have been suggested not to be concentrated at the hinge points and the MHP and DLHPs to be resistant to cytochalasins (Smith and Schoenwolf, 1988; Ybot-Gonzalez and Copp, 1999). Most of the mutants in genes related to actin cytoskeletons have defects only of cranial neurulation, but not in spinal neurulation (Copp et al., 2003). NUAK1/2 might target a novel component in which loss results in inefficient apico-basal elongation and apical constriction of neuroepithelial cells or the lack of DLHP formation, and the elucidation of the NUAK1/2 primary targets should expand our understanding of the molecular mechanisms of cytoskeletal reorganization in neuroepithelial cells during neurulation.


Generation of NUAK1 and NUAK2 Mutant Mice

NUAK1 mutant mice (Acc. No. CDB0029K; were generated as previously described (Hirano et al., 2006). The targeting vector to generate NUAK2 mutant mice (Acc. No. CDB0065K; was constructed as described ( A BAC clone (RP23-423B16), which contains the translational start site of NUAK2 gene, was obtained from BACPAC Resources. To amplify the 7.3-kb 5′ long arm by PCR, the 23-bp (5′-GCCAGCACAGGCTGCAGCAGAGA-3′) and 27-bp (5′- GGCAGCAGTAGCAGCGCGCGCTGAGCC-3′) and to amplify the 2.0-kb 3′ short arm, the 24-bp (5′-CTTGCGGTGTCCTCGGCTGCGTTG-3′) and 24-bp (5′- GGGAGCTAGGAGACTCACAGGAGG -3′) were used as primers, respectively. Each product was cloned into DT-A-pA/lox71/LacZ-pA/frt/PGK-Neo/frt/loxP/pA vector ( at SalI/NotI (5′ arm) or SwaI/XhoI sites (3′ armt). The targeting vector was electroporated into TT2 ES cells (Yagi et al., 1993) and 128 G418 resistant clones were screened by PCR for homologous recombinants. Primers used to detect them as a 2.7-kbp product were: forward primer, 5′-GTACTCGGATGGAAGCCGGTCTTGTC-3′; reverse primer, 5′-AGACCAATGGCTTCGGTGAGACCC-3′. Five clones were PCR positive, and confirmed as homologous recombinants by Southern blot analysis. The homologous recombinant TT2 ES cells were injected into 8-cell-stage ICR embryos to generate germ-line chimera ( Chimeric male mice were mated with C57BL/6 females to generate F1 heterozygotes. Genotypes of mice or embryos were routinely determined by PCR with genomic DNAs from tail: wild allele was detected as a 597-bp band by primer P1 (5′-ACGTTTTATAGCAATTCGGAACG) and primer P2 (5′-TCCTTCCATTCTACTCTGTTGACG), and mutant allele as a 380-bp band by primer P1 and primer P3 (5′- CATTCGCCATTCAGGCTGCGC). The PCR consisted of: pretreatment at 94°C for 2 min, 40 cycles of the PCR (denaturation for 30 sec at 94°C, annealing for 30 sec at 62°C, annealing and extension for 90 sec at 72°C), post-treatment at 72°C for 7 min. Mice were generated and housed in environmentally controlled rooms under the RIKEN Center for Developmental Biology (CDB) guidelines for animal and recombinant DNA experiments.


RT-PCR was conducted as described (Hirano et al., 2006). Sizes of the products and primers used were: GAPDH, 305 bp with forward (5′-TGTCATCAACGGGAAGCCCA-3′) and reverse (5′-TTGTCATGGATGACCTTGGC-3′); NUAK2, 278 bp with forward (5′-CCACATCATTGCCATCCATG-3′) in exon 2 and reverse (5′-CTTTGTGGTACAGGTTGGAG-3′) in exon 5.

Western Blot Analysis

To produce an antibody against NUAK2 (492–640 aa), the relevant DNA was obtained by PCR using primer F (5′-GGGGGAATTCCATGTT CTCCCGCACAGCCTTAGAAGG) and primer R (5′-GGGGCTCGAGCCTCAGCTGAGCTTTGAGCAGATTCC). The DNA fragment was inserted into pGEX-4T3 vector to generate a protein for immunization to mice. Antibodies against MLC2 (no. 3672) and pMLC2 (no. 3671) were obtained from Cell Signaling (Danvers, MA). Western blotting was performed as described (

Histological Analysis

Mouse embryos were fixed with 4% paraformaldehyde at 4°C or Bouin's fixative solution at room temperature for 18–24 hr. Specimens were dehydrated and embedded in paraffin. Serial sections (thickness of 10 μm) were prepared and stained with hematoxylin and eosin.

βGal Expression

βGal expression was determined as described (Kimura et al., 1997, 2000). Embryos were dissected in phosphate-buffered saline (PBS), fixed in 0.2% glutaraldehyde, 0.02%Nonidet-P 40 in PBS at room temperature for 5 min, washed with PBS twice for 5 min, dipped into staining solution (1 mg/ml X-gal, 2 mM MgCl2, 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6 in PBS), and incubated at 37°C overnight.

RNA In Situ Hybridization

Embryos were dissected in PBS and fixed in 4% paraformaldehyde at 4°C for 18–22 hr. Whole mount and section in situ hybridizations were performed with digoxigenin-conjugated probes as described (Wilkinson and Nieto, 1993). The probes used were: NUAK1/Omphk1, NUAK2/Omphk2 (Hirano et al., 2006), Uncx4.1 (MGC5718115), twist1 (MGC6516673), Cerl1 (Belo et al., 1997), noggin (McMahon et al., 1998), Msx1 (Hill et al., 1989), Wnt3a (Roelink and Nusse, 1991), Shh (Echelard et al., 1993) and BMP2 (Dickinson et al., 1990). Crabp1 probe was a kind gift from Dr. Bernhard G. Herrmann (MAMEP database,


Antibodies used for immunohistochemistry were: ZO1 (Invitrogen, Carlsbad, CA; 339100), phospho- PKCζ/λ(Thr 410/403) (Cell Signaling, Danvers, MA; no. 9378), PALS1 (Millipore, Billerica, MA; 07-708), N-Cadherin (BD Transduction Laboratories, San Diego, CA; 610920), Laminin (Harbor Bio-Products, Norwood, MA; 5008), β1 Integrin (Millipore, no. MAB1997), cortactin (Cell Signaling, 3502), Nonmuscle Myosin Heavy Chain II-B (Covance, Princeton, NJ; PRB-445P), Phospho-Myosin Light Chain 2 (Cell Signaling, no. 3671), Myosin Light Chain 2 (Cell Signaling, no. 3672), α-Tubulin (Sigma, St. Louis, MO; T9026), γ-Tubulin (Sigma, T5192), Acetyl-α-Tubulin (Lys40) (Sigma, T6793), MAP2 (Sigma, ·4403), phospho-Histone H3 (Ser10) (Upstate, East Syracuse, NY; no. 06-570), and Cleaved Caspase-3 (Asp175) (Cell Signaling, no. 9661). Antibodies against Shroom3 and RhoA were kindly gifted by Dr. M. Takeichi (Nishimura and Takeichi, 2008) and Dr. S. Yonemura (Yonemura et al., 2004), respectively. Phallotoxin to stain F-actin was purchased from Invitrogen (R415). Embryos were fixed and embedded in 66% OCT compound, 6.6% sucrose in PBS. Serial sections (thickness of 12 μm) were prepared and rehydrated. After blocking with 3% BSA in tris-buffered saline containing 0.1% Tween 20 (TBS-T), specimens were incubated with primary antibodies at 4°C overnight, washed three times with TBS-T for 5 min at room temperature, and treated with AlexaFluor 488-, 555-, 594-, or 647-conjugated secondary antibodies (Invitrogen). For whole mount F-actin staining (Fig. 7Ba–f), embryos were fixed with 4% paraformaldehyde at 4°C for 18–24 hr, blocked with 10% sheep serum in TBS-T for 2 hr, stained with Phalloidin (Sigma P1951) at 4°C for 18–24 hr, washed four times with TBS-T at 4°C for 4 hr, embedded with 1% low-melting point agarose in PBS, and put on a 35-mm glass bottom culture dish. The samples were imaged by a Nikon A1 confocal microscope with 40× objective lens.


We appreciate the critical discussion and insight provided by Dr. T. Terashima. We are also grateful to the Laboratory for Animal Resource and Genetic Engineering for mutant mouse production and the housing of mice.