|Home | About | Journals | Submit | Contact Us | Français|
Myosin-X (MyoX) belongs to a large family of unconventional, non-muscle, actin-dependent motor proteins. We show that MyoX is predominantly expressed in cranial neural crest (CNC) cells in embryos of Xenopus laevis and is required for head and jaw cartilage development. Knockdown of MyoX expression using antisense morpholino oligonucleotides resulted in retarded migration of CNC cells into the pharyngeal arches, leading to subsequent hypoplasia of cartilage and inhibited outgrowth of the CNC-derived trigeminal nerve. In vitro migration assays on fibronectin using explanted CNC cells showed significant inhibition of filopodia formation, cell attachment, spreading and migration, accompanied by disruption of the actin cytoskeleton. These data support the conclusion that MyoX has an essential function in CNC migration in the vertebrate embryo.
The neural crest (NC) originates at the boundary between the neural and non-neural ectoderm as the result of complex inductive interaction (Huang and Saint-Jeannet, 2004). As the neural folds meet and fuse, NC cells undergo a classical epithelial-mesenchymal transition, delaminate from the neuroepithelium and subsequently migrate throughout the body, finally differentiating into a large number of distinct cell and tissue types (Bronner-Fraser, 1993; Le Douarin, 1999). In the anterior region of the embryo, the cranial NC (CNC) cells originating from the presumptive hindbrain migrate in coherent streams into the ventrally located pharyngeal pouches, giving rise to the pharyngeal arches that differentiate into jaw cartilage and bone, as well as some other structures including portions of the cranial ganglia and nerves (Schlosser, 2006). One key transcription factor is TFAP2a, one of a five-member subfamily of transcriptional activator (Eckert et al., 2005). TFAP2a has been shown by loss of function experiments to be essential for the maintenance and subsequent differentiation of CNC in vertebrates (Schorle et al., 1996; Zhang et al., 1996; Knight et al., 2003; Luo et al., 2003). In neuralized Xenopus ectoderm, ectopic expression of TFAP2a resulted in repression of neural genes and activation of the NC transcriptional program (Luo et al., 2003). We have used this as an assay for identification of TFAP2a target genes expressed in NC (Luo et al., 2005). Among the most highly induced genes discovered in this screen was the unconventional myosin, MyoX.
MyoX is one of approximately a dozen different classes of myosin expressed in vertebrates (Berg et al., 2001) characterized by a N-terminal actin-dependent motor domain and variable C-terminal tails with binding sites for a number of cytoskeletal components, plasma membrane, signaling molecules and other factors (Sousa and Cheney, 2005). The head of MyoX binds directly to actin filaments and generates intracellular movement (Homma et al., 2001). The MyoX tail includes a FERM domain (band 4.1,ezrin,radixin and moesin) which interacts with some beta integrins, and is important in cell attachment (Zhang et al., 2004). In cultured cells, MyoX has been shown to be both necessary and sufficient for the formation of filopodia (Bohil et al., 2006). MyoX has additional functions in phagocytosis (Cox et al., 2002), intercellular adhesion (Yonezawa et al., 2003) and endocytosis (Tacon et al., 2004). However, the functions of MyoX in intact organisms have not been thoroughly studied. To investigate this in vertebrate development, we have used loss-of-function analysis in Xenopus embryos. We have found that MyoX is strongly expressed in the CNC, and that the zygotic expression of MyoX is required for CNC cell migration from the hindbrain to the pharyngeal arches. Thus MyoX has been established as an essential component of craniofacial development in the vertebrate embryo.
MyoX is vertebrate-specific and has been shown to be expressed in many cells and tissues (Berg et al., 2000). MyoX mRNA is present maternally, and at all stages of development in Xenopus (Fig. 1A). MyoX RNA is not uniformly expressed spatially, however, and is most abundant in cranial and trunk NC at neurula stages (Fig. 1B-D). MyoX RNA is also abundant in paraxial mesoderm and the anterior neural plate (Fig. 1B). By tailbud stage (st. 21) MyoX expression persists in migratory CNC cells, then declines following the completion of NC migration to form the pharyngeal arches (Fig. 1E,F). By stage 33/34 expression in cranial ganglia is visible while pharyngeal arch expression becomes undetectable (Fig. 1G, H).
Maternal MyoX has been shown to have critical functions in both meiosis and mitosis in Xenopus (Weber et al., 2004; Woolner et al., 2008). In order to test if MyoX is required for CNC development, we used antisense morpholino oligonucleotides (MOs) to specifically inhibit zygotic MyoX pre-mRNA splicing. This strategy circumvents inhibition of the essential maternal MyoX expression, allowing embryonic survival beyond early cleavage stages. Two MOs were designed (named MO1 and MO2), and both were shown to specifically and substantially block processing of MyoX RNA (see Supporting Information, Fig.S1 online). Injection of either MO resulted in significant reductions in head size, (100 %, n=255 for MO1; 93%, n=231 for MO2; Fig. 2A). Alcian blue staining revealed that all head cartilage elements were present in morphants, but were dramatically reduced in size (Fig. 2B).
To quantify this defect, we measured tadpole head width following injection of MO or water as control. As shown in Fig.2C, the average width of controls was 1.60 mm (blue dotted line), whereas that of MO1 or MO2 was 1.13 mm. We also performed the MO co-injection along with increasing doses of GFP-fused MyoX mRNA in an attempt to rescue the morphant phenotype and rule out nonspecific MO effects. At the 6ng dose of co-injected GFP-MyoX RNA, the narrow head phenotype was rescued significantly, especially for MO2, which approached control levels. This strongly supports the conclusion that MyoX is required for head cartilage development.
During Xenopus head development, CNC cells migrate from hindbrain in a segmental manner to contribute to the skeletal cartilage (Sadaghiani and Thiebaud, 1987). Induction of NC occurs during gastrulation as the result of complex interactions involving neural ectoderm and adjacent epidermis and paraxial mesoderm (Bonstein et al., 1998). As shown in Figure 1, MyoX is expressed at high levels in the latter tissue, raising the possibility that the observed CNC defects might be the result of abnormal NC induction due to defects in adjacent tissues, such as mesoderm or epidermis. This was investigated by in situ hybridization with tissue-specific probes at early embryonic stages. No significant changes were observed in MyoX knockdown embryos (Fig. 3A), including snail2 (also called slug), which is an early marker for NC induction (Mayor et al., 1995). In addition, gastrulation proceeded normally in these embryos (data not shown). Therefore the observed defects in craniofacial development must be the result of abnormalities in subsequent migration or defective terminal differentiation of NC cells. This interpretation was supported by in situ hybridization with Sox9 (Spokony et al., 2002), and Sox10 (Aoki et al., 2003) probes at tailbud stages, both of which clearly demarcate NC. Both markers revealed retarded migration of cranial NC cells on the MO-knockdown side (Fig. 3B). The migration defect was not accompanied by apoptotic cell death: TUNEL assays from early neurula to st. 24 tailbud did not show significant increases in signal on the MO-knockdown side (Fig. S2). The morphant phenotype was not limited to cartilage, but also affects other CNC derivatives in the head. As shown in Figure 3C, MyoX knockdown also inhibited trigeminal nerve outgrowth, indicated by Sox10 expression at st. 32 (Cheng et al., 2000; De Bellard et al., 2002).
The transplantation assay has been well established to study the migratory behavior of CNC cells in Xenopus (Borchers et al., 2000). To test for cell autonomy of the MyoX knockdown phenotype, transplantation experiments were carried out. Premigratory CNC, with overlying ectoderm, was isolated from fluorescently-labeled MyoX knockdown and control embryos and transplanted into unlabeled host embryos. To lessen the possibility of MyoX knockdown in adjacent tissues, which might affect the CNC phenotype, the MO in these experiments was targeted to the appropriate blastomeres at the 16–32 cell stage. This strategy would not prevent possible knockdown of MyoX activity in adjacent epidermis, but the MyoX RNA expression level is extremely low in epidermis (Fig. 1C), and this tissue developed normally in the knockdown embryos (Fig 3A). Therefore we believe a role for epidermal MyoX in the CNC phenotype is unlikely. As shown in Figure 4, control transplants migrated normally and formed essentially complete craniofacial structures by St. 45. However, MO2 knockdown cells exhibited delayed migration and significant reduction in cranial cartilage on the grafted side, similar to the previous experiments. These results support the conclusion that MyoX functions in NC migration, and does so in a cell-autonomous manner.
Xenopus laevis CNC cells migrate efficiently in vitro on a fibronectin substrate, requiring integrinα5β1 as the primary receptor complex (Alfandari et al., 2003). To test the ability of MyoX knockdown cells to migrate in vitro, CNC explants were dissected from stage 17 embryos injected with either control MO or MO2 and plated on plastic dishes coated with fibronectin. Migration was monitored over a 20-hour period. As shown in Fig. 5A and Movie S1, control NC cells migrated extensively from the explant, often maintaining coherent streams that likely correspond to pharyngeal arch populations in vivo. However, the majority of CNC cells from MO2-injected embryos failed to migrate. Furthermore, knockdown cells exhibited a rounded morphology lacking filopodia (Fig. 5D, E, Movie S2). These cells attached only weakly to the substrate and, unlike control NC cells, could be easily stripped from the slide by gentle pipetting or agitation (data not shown). This phenotype was more severe than what was observed in the intact embryo. However, since these defects were significantly rescued by co-injection of MyoX RNA, nonspecific artifacts can be ruled out, supporting the conclusion that MyoX is required for CNC cell migration, and probably for substrate adhesion as well, at least in vitro (Fig. 5C, D, E and Movie S3).
MyoX is a ubiquitous, essential protein required for basic cellular processes including mitosis, and meiosis (Weber et al., 2004; Bohil et al., 2006; Woolner et al., 2008). In addition to these activities, the MyoX gene has tissue-specific functions, including cargo transport in neurons and endothelial cells (Pi et al., 2007; Zhu et al., 2007). In this report, we have shown that MyoX is expressed at relatively high levels in the CNC, and that this expression is critical for the migration and subsequent differentiation of CNC into derivative tissues such as cranial cartilage and ganglia.
The critical function of MyoX in filopodial formation has been well-established (Bohil et al., 2006). At the leading edge of migratory cells filopodia can act as sensors to explore the local environment for directional cues. Filopodia also provide initial attachment to the substrate (Partridge and Marcantonio, 2006) which, when stabilized, provides the traction enabling the cell to move forward (Heidemann et al., 1990; Bridgman et al., 2001). Filopodial formation is dependent on MyoX and this is likely to be an essential component of NC cell migration. One possible explanation for the craniofacial phenotype caused by knockdown of MyoX is that reduced CNC cell migration resulted in a deficit of pharyngeal arch mesenchyme. Another alternative is suggested by our observation that MyoX knockdown CNC cells were unable to attach to fibronectin in vitro. If a similar defect occurs in vivo in the MyoX knockdown CNC cells, this might result in signaling defects. For example, in mammalian NC cells association of integrin with extracellular matrix components has been shown to activate multiple signal transduction responses (Desban et al., 2006). Since NC cells acquire much of their differentiation potential by inductive interactions during and after migration (Le Douarin et al., 2004), it is possible that MyoX knockdown CNC cells are unable to receive such inductive cues, leading to a failure to differentiate properly. On the other hand, while slowed, CNC migration in MyoX knockdown embryos did not appear to be mis-directed; cells derived from different regions from the hindbrain targeted the appropriate pharyngeal arches, and other destinations in the head. Furthermore there was little if any increase in apoptosis resulting from MyoX knockdown, suggesting that at least some signaling between CNC and adjacent cells was intact. Nevertheless we can not rule out the possibility that separate signals could support migration targeting versus differentiation cues. Conceivably, the role of MyoX in CNC could be even more complex. In addition to integrins, other cargoes for MyoX have been identified, such as DCC (Zhu et al., 2007), VASP (Tokuo and Ikebe, 2004), and the receptor for BMP6 (Pi et al., 2007). Defects in transport of such molecules could result in abnormal CNC migration or differentiation.
Loss of maternal MyoX results in disruption of cleavage and embryonic lethality (Woolner et al., 2008); YS Hwang, unpublished). Based on the absence of apoptosis in the knockdown embryos (Fig. S2) and the absence of alterations in early gene expression in NC or other cell types, we speculate that in contrast to maternal MyoX, the zygotic expression of this gene is not required for normal cell division. One trivial alternative could be that the knockdown of MyoX in our experiments was incomplete, providing enough residual protein for normal spindle formation. It is also conceivable that another member of the myosin gene family could provide some redundancy for essential MyoX functions, although there is no evidence to support this notion. A more interesting possibility is that the requirement for maternal MyoX in mitosis is specific to the rapid cleavage cycles during blastula stages in Xenopus (Newport and Kirschner, 1984) and is dispensable in more slowly dividing cells. Additional loss-of-function experiments will be required to resolve this question.
MyoX expression is also elevated in other tissues in early development, such as trunk NC, paraxial mesoderm forebrain and presumptive gut endoderm (not shown). While we have not examined derivatives of these tissues in detail, they did not exhibit any obvious defect except for the gut, which exhibited defective coiling (not shown). Forebrain abnormalities have not been ruled out, but would be difficult to discriminate from secondary effects of the cranial bone deficiency. The most likely interpretation is that CNC cells have an elevated sensitivity to loss of MyoX compared to the other highly-expressing tissues. Presumably this differential sensitivity reflects some special property of CNC, such as the highly migratory nature of this cell type, necessitating rapid and dynamic transport of adhesion proteins including integrins to filopodial tips where cell-substrate and cell-cell contacts are promoted. The observed defect in projection of cranial nerve V could also be an indirect result of delayed NC migration. However, an alternative hypothesis is that MyoX is required for outgrowth of the cranial nerve per se, similar to what has been reported for neurons from both chick and mouse embryos in which MyoX function has been inhibited (Zhu et al., 2007).
Compared to in vivo migration defects, the phenotype of isolated MyoX knockdown CNC explants on fibronectin substrate was very severe. Not only was migration virtually eliminated, but cell-substrate adhesion was severely reduced, and cytoskeletal organization was disrupted as well. One possibility is that in vivo, other extracellular matrix components in addition to fibronectin might contribute to providing the substrate for CNC migration. In Xenopus, integrinα5β1 and α3β1 are present on the CNC cell surface (Alfandari et al., 2003). Since integrinα3β1 binding to laminin 5 is involved in cell migration and cell invasion (Giannelli et al., 2002), it is possible this type of laminin also important for CNC migration in vivo. It is also possible that the artificial environment provided by purified fibronectin on glass or plastic surfaces induces additional stress to CNC cells, or lacks signals that are required for robust migration in vivo.
We have shown that MyoX has been recruited to serve as an essential component of CNC cell adhesion and migration. NC cells are among the most migratory in development, needing to rapidly travel from their origin in the neural folds throughout the entire body, following defined paths and undergoing complex terminal differentiation depending on signals received en route. Compared to the initial induction of NC, such later differentiation signals are very poorly understood. Understanding the role of integrin-dependent signaling in CNC, identifying other proteins in addition to integrinβ1 with which MyoX interacts in CNC, and further elucidating the importance of MyoX-dependent filopodial formation in CNC development will be important areas for future research.
Myc-MyoX was generated by transferring a PCR-amplified MyoX coding fragment into a modified pCS2+ vector that adds a single myc epitope at the amino-terminus. GFP-MyoX was a gift from William M. Bement (Weber et al., 2004).
Based on Xenopus tropicalis genome information (Ensembl Gene ID: ENSXETG00000007165), five short intronic regions within the MyoX motor domain were selected as candidate sites for splice-inhibitory morpholino design. Primers (sequences available upon request) flanking these introns in the X. laevis MyoX mRNA sequence were used to amplify X. laevis genomic DNA fragments, which were cloned into pGEM-T (Promega) and sequenced. One splice acceptor site (intron 3, MO1) and one splice donor site (intron9, MO2) were chosen as targets and were synthesized by Gene tools, LLC. MO1: TCC CCA CTG TAA CGG GAA GCA TGG T. MO2: GGT GCT GCT GCA TTA CCT GCT TTG C. MO doses were determined by titration using cranial morphology and RT-PCR to evaluate effectiveness; a minimum of 60 ng/embryo of each MO was required.
Xenopus embryos were obtained and staged according to standard procedures (Sive et al., 2000). For RT-PCR, total RNA was extracted from embryos using TRIzol (Invitrogen), and synthesis of oligo-dT primed cDNA was performed with Superscript II reverse transcriptase (Invitrogen) according to manufacturer's instructions. For the MyoX in situ probe a fragment corresponding to the FERM domain and 3’UTR was amplified by PCR from the NIBB clone XL085p13 and cloned in pBluescript SK (−). Other antisense probes for Snail2 (Mayor et al., 1995), Sox2 (Mizuseki et al., 1998), Sox9 (Spokony et al., 2002), Sox10 (Aoki et al., 2003), MyoD (Frank and Harland, 1991), XK81 (Jonas et al., 1985) were labeled with digoxigenin, and whole-mount in situ hybridization procedures were performed according to standard methods (Harland, 1991). Alcian Blue cartilage staining was performed as described (Pasqualetti et al., 2000). TUNEL staining was previously described (Huang et al., 2007) using 1X TdT buffer (TdT recombinant; Invitrogen; 10533–065), 1μM digoxigenin – dUTP (Roche; 11-093-088-910) and anti-digoxygenin AP antibody (Roche; 11-093-274-910).
CNC transplantation (Borchers et al., 2000) and in vitro migration assays (DeSimone, 2005) were performed as previously described. Images were captured at 8m 40s intervals for 26 hr using an automated stage-aided Zeiss Duoscan confocal microscope (10x objective). The area of the migrated cell mass was measured using the Metamorph program. Data sets from two separate experiments were combined to generate plots and to perform ANOVA (analysis of variance).
CNC explants were cultured for 20 hr at 19°C in a plastic chamber slide, then fixed by gently adding an equal volume of 2XFG fix (1X FG: 3.7% formaldehyde, 0.25% glutaraldehyde, 0.2% TritonX-100 in Pipes buffer; (Gard et al., 1997), after 5 min samples were washed 3 times for 30 min in 1XPBS, 0.1% Tween20, treated with rhodamine-labeled phalloidin (Invitrogen Molecular Probes, Cat#R415, 5u/ml) at RT for 2 hr, then washed 3 times for 30 min in 1xPBS. Slides were mounted in VECTASHIELD with DAPI (H-1200), and air-dried in dark overnight. Images were obtained using a Zeiss Axioplan fluorescent microscope.
To estimate the significance of differential migration and cell rounding, we carried out analysis of variance (ANOVA; Fig.5C, E) with Dunnett T3 Post Hoc Tests and Duncan’s homogenous subset grouping (Fig. S4). Data from the four experimental groups representing differential cellular level of MyoX underwent One-way ANOVA analysis. The effects of four different conditions on CNC cell migration and attachment were highly significant (sig. < 0.000~) showing that the MO2 injected group was the most affected and significantly different from the other three groups. (Duncan’s Homogenous subset grouping; Fig. S4).
Movie S1-S4: Movies from in vitro migration assays (Fig.5A). Images were captured from 1 hour after all the explants were plated, at 8m 40s intervals for 26 hr resulting in a total of 180 still pictures. These four 18-second-long movies contain 26 hour events, in which 1 second roughly represents 1 hour 30 minutes. For details, see Experimental Procedures section.
Movie S1 cMO group CNC cells, one example from a total of 20 movies. Control NC cells located at the base of the explant migrated out first, followed by upper tiers of cells. For up to 9 hr explants often maintained some segmental movement. After 13 hr of culture a large proportion of cells began to exhibit rounded morphology.
Movie S2 MO2 group CNC cells, one example from a total of 19 movies. The majority of CNC cells from MO2-injected embryos exhibited a rounded morphology from the onset of the culture period. These cells failed to attach or spread on the fibronectin substrate, and could be observed being moved by contact with the minority of cells that attached and migrated, albeit poorly, from the explant. By 20 hr of culture essentially all cells exhibited the rounded, poorly attached phenotype.
Movie S3 MO2 + MyoX, rescue group CNC cells, one example from a total of 20 movies. The severe defects in cell attachment, spreading and migration from MO2-injected embryos were significantly rescued by co-injection of MyoX RNA.
Movie S4 MyoX, RNA alone group CNC cells, one example from a total of 12 movies. Migration from CNC explant cells from MyoX RNA-injected embryos was essentially normal, although there was a small but significant decrease observed in the proportion of rounded cells compared to controls (Fig.5E).
(A) X. laevis MyoX PCR primers were designed for amplification of intron sequences, based on genome data from X. tropicalis. One splice acceptor site and one splice donor site corresponding to introns 3 and 9, respectively, were selected for MO design (MO1 and MO2). A control MO, with no homology to MyoX, was also obtained. (B) Two-cell-stage embryos were injected into both blastomeres with a total of 100, 80 or 60 ng of each MO, cultured and harvested at st.17, followed by RT-PCR using intron-spanning primers (primer 1F/R and 2F/R). The positions of unspliced and correctly-spliced PCR products are indicated to the right. MO1 blocked splicing almost completely, while MO2 was somewhat less efficient, but both were highly specific. In both instances, the splice disruption would lead to premature termination of translation within 15 amino acids (not shown).
Embryos injected into one cell at the two-cell stage with control MO (cMO) or MO2, along with cytoplasmicβ-galactosidase RNA (Detrick et al., 1990) for lineage tracing, followed by whole mount TUNEL assay on embryos at various stages. There was no statistically significant difference between injected and uninjected sides, or between cMO and MO2 samples. The TUNEL+ cells could be distinguished from β-galactosidase staining by viewing in a dissecting microscope. A higher magnification of the injected side at stage 24 is shown, with TUNEL+ cells indicated with red arrowheads. Unpaired Student t-test was performed to determine statistical significance. Data are summarized by the whisker graphs at the right, the black dots indicate the interval that included 95% of data points.
(A-C) Expression of MyoX at St.31 in cranial nerve V and VII. (F) Cranial nerve VII starts branching at St.32 (purple arrow). (I, L) The ascending (red arrow) and descending (blue arrow) branches of VII become prominent at St.35. The ascending branch approaches the anterior-dorsal aspect of the cement gland and reaches there at St.38 MyoX expression demarcates nerve VII at this stage. (D, G) Somites express MyoX throughout tail bud stages. (M-P) MyoX is expressed in all brain regions. For anatomy of cranial nerves and brain see (Honore and Hemmati-Brivanlou, 1996; Nagata et al., 2003; Huang et al., 2007). F: forebrain, M: midbrain, H: hindbrain.
To evaluate startistical significance of the data in Figure 5, the Dunnett T3 Post Hoc test and Homogeneous Subgrouping was performed in addition to one-way ANOVA test using the SPSS (ver.16) program. These tests confirm that in the area ratio and percentage of round cell comparisons, the MO2 group was different from the control, rescue and RNA alone groups (MyoX). The rescue group was different from both control and MO2 group in both aspects. RNA alone group was slightly but significantly different from both control and MO2 groups in the area ratio measurement, but not in the round cell percentage measurement. The confidence for all differences was at the 95% level (subset for alpha = 0.05).
We thank Dr. William Bement for providing the GFP-MyoX construct. We are also grateful to V. Schram (Microscopy and Imaging Core, NICHD) for valuable assistance with confocal microscopy, and Dr. Sheran Law for discussions. This work was supported by the Intramural Research Program of the National Institute of Child Health and Human Development, NIH.