Myristoylated alanine-rich C kinase substrate (MARCKS) is an actin-binding, membrane-associated protein expressed during Xenopus embryogenesis. We analyzed its function in cytoskeletal regulation during gastrulation. Here, we show that blockade of its function impaired morphogenetic movements, including convergent extension. MARCKS was required for control of cell morphology, motility, adhesion, protrusive activity, and cortical actin formation in embryonic cells. We also demonstrate that the noncanonical Wnt pathway promotes the formation of lamellipodia- and filopodia-like protrusions and that MARCKS is necessary for this activity. These findings show that MARCKS regulates the cortical actin formation that is requisite for dynamic morphogenetic movements.
During Xenopus gastrulation, mesoderm migrates to the inside of the embryo and moves along the blastocoel roof to establish the three germ layer structure. This process involves several morphogenetic cell movements including mesendoderm extension and convergent extension. During mesendoderm extension, cells migrate along the blastocoel roof in contact with fibronectin (FN) fibrils (Winklbauer, 1990; Davidson et al., 2002). In convergent extension, cells are polarized and elongated mediolaterally, then the cells are intercalated. This movement forms the dorsal mesodermal structure and extends the anteroposterior body axis (Shih and Keller, 1992; Wallingford et al., 2002). The noncanonical Wnt pathway has been implicated in the regulation of convergent extension (Kuhl, 2002; Tada et al., 2002). One of the intracellular signaling components, Xenopus Dishevelled (Xdsh), plays a pivotal role in this process. When the function of Xdsh is inhibited, the polarity of the mesodermal cells is not established normally (Wallingford et al., 2000).
Because these cell movements are accompanied by dynamic changes in cell polarity, morphology, and motility, it is very likely that cytoskeletal dynamics are carefully regulated. Thus, we sought to analyze the regulatory mechanism of cytoskeletal dynamics during gastrulation. We decided to focus on myristoylated alanine-rich C kinase substrate (MARCKS). Mammalian MARCKS has been shown to interact with actin (Arbuzova et al., 2002). It has been reported that Xenopus MARCKS is expressed maternally and throughout embryogenesis (Ali et al., 1997; Shi et al., 1997), but its role in development was not well understood.
Here, we report that the loss of MARCKS function severely impaired gastrulation movements. MARCKS regulates the cortical actin formation, cell adhesion, protrusive activity, and cell polarity control during gastrulation. We further show that MARCKS is necessary for the protrusive activity regulated by the noncanonical Wnt pathway. These findings show that MARCKS regulates the cortical actin formation that is requisite for dynamic morphogenetic movements.
Results And Discussion
To investigate the function of MARCKS in Xenopus development, we conducted loss of function experiments using antisense Morpholino oligonucleotides (Mo). First, we examined the specificity of MARCKS Mo (Fig. S1). The Mo specifically and effectively inhibited epitope-tagged MARCKS protein synthesis, leading us to expect that it could inhibit the endogenous MARCKS protein synthesis.
Using MARCKS Mo, we analyzed MARCKS function in development. When it was injected into the dorsal marginal zone (DMZ) of four-cell embryos, the embryos showed a gastrulation-defective phenotype (Fig. 1 A). The involution of the mesoderm was impaired and the blastopore remained open. A similar phenotype was observed when MARCKS mRNA was injected. The phenotype of MARCKS Mo was partially rescued by coinjection of MARCKS mRNA (Fig. 1 B). The rescue was imperfect probably because MARCKS overexpression also inhibited gastrulation movements. As discussed below, however, cell biological effects of MARCKS Mo were efficiently rescued by MARCKS mRNA. Over- and under-expression of MARCKS may have opposite effects at a cellular level, but both of these effects may negatively influence gastrulation movements. MARCKS is essential for gastrulation and its level must be tightly regulated.
It has been reported that MARCKS-like protein (XMLP) is also expressed in Xenopus embryo (Zhao et al., 2001). Although XMLP is similar to MARCKS (23% amino acid identity), XMLP-Mo injected embryos showed malformations of the anterior axis and eye defect, but the gastrulation defect was not reported (Zhao et al., 2001). They seem to play distinct roles in Xenopus development.
To determine whether this gastrulation defect was caused by a defect in mesodermal differentiation, we examined the expression of the dorsal mesodermal markers. At the gastrula stage, MARCKS Mo-injected embryos expressed chordin at the same level as control embryos (Fig. 1 C). In tadpoles, the notochord and somites were formed in the MARCKS Mo-injected embryos, but the extension of these tissues was severely inhibited (Fig. 1 D). We also tested the expression of the mesodermal markers in DMZ explants by RT-PCR (Fig. 1 E). The expression of these markers was not inhibited by MARCKS Mo. These results indicated that the phenotype was caused, not by a defect in mesoderm differentiation, but by a defect in morphogenetic movements.
Next, we tested whether the loss of MARCKS function affects the animal cap elongation, which mimics convergent extension movements during gastrulation (Fig. 1 F). MARCKS Mo blocked the elongation by activin, and it was rescued by coinjecting MARCKS mRNA without the UTR, suggesting that MARCKS is required for convergent extension.
During mesodermal convergent extension, the cells become polarized, align mediolaterally, and are then intercalated. To test how MARCKS is involved in this process, the convergent extension in DMZ explants was observed microscopically. MARCKS Mo, Rhodamine dextran, and mRNA encoding membrane-binding Venus (mb-Venus) were coinjected into one of the two dorsal blastomeres (Fig. 2 A). As a control, mb-Venus mRNA alone was injected into the other dorsal blastomere. At the gastrula stage, the DMZ explants were isolated and cultured on a cover glass coated with FN. These explants adhered to the FN, and convergent extension movements occurred subsequently in the mesoderm (Kinoshita et al., 2003). In the absence of MARCKS Mo, red and nonred cells were polarized and intercalated. In the MARCKS Mo-injected explants, the nonred cells, which were assumed to lack the Mo, were polarized and showed convergent extension. In contrast, the red cells (Mo-injected cells) were not polarized and did not participate in the intercalation. Thus, MARCKS is essential for the cell polarization and movement during convergent extension.
In addition to convergent extension, an important mechanism regulating gastrulation movements is mesendoderm extension (Davidson et al., 2002). To test whether MARCKS is required for this process, DMZ explants were cultured on FN-coated dishes according to the method developed by Davidson et al. (2002). Mesendodermal cells migrated on the FN substrate as an intact mantle (Fig. 2 B). When Venus mRNA and the control Mo were coinjected, the Venus-expressing cells dispersed broadly, and some cells migrated to the front. In contrast, MARCKS Mo-injected cells rarely migrated on the FN substrate. We examined 15 explants and confirmed that none of the MARCKS Mo-injected cells reached the leading edge of the migrating mesendoderm. This indicated that MARCKS was required for mesendoderm extension as well as convergent extension.
We examined whether MARCKS Mo affects the adhesion to FN of cells dissociated from the DMZ explants. MARCKS Mo and Venus mRNA (green) or Rhodamine dextran (red) was injected dorsally (Fig. 2 C). DMZ explants were isolated and dissociated in Ca2+-Mg2+–free medium. Dissociated cells were cultured on FN-coated dishes, and cells that adhered to the dish were counted (Fig. 2 C). When MARCKS Mo was coinjected with Venus, the adherence of Venus-expressing cells was extremely reduced. In contrast, when MARCKS Mo was coinjected with Rhodamine dextran, these red cells rarely adhered to the dish. A few cells containing MARCKS Mo were found on the dish, but these cells were rounded up and did not spread out on the dish. This indicates that MARCKS is essential for cell adhesion and spreading on FN. The effect of MARCKS Mo on the cell adhesion was rescued by coinjection with MARCKS mRNA (unpublished data).
Next, we tested whether MARCKS Mo affected the protrusive activity in mesodermal cells. Mesodermal cells had many filopodia-like protrusions when the DMZ explants adhered to a FN-coated dish (Fig. 2 D). MARCKS Mo severely reduced the number and the length of these protrusions. The effect of MARCKS Mo on the protrusive activity was rescued by coinjection with MARCKS mRNA. Thus, MARCKS is required for the protrusive activity, which may directly correlate with the control of cell adhesion and motility. The inhibition of cell adhesion and migration on the FN fibrils that cover the blastocoel roof may contribute to the gastrulation defect caused by MARCKS Mo.
The actin-binding domain of mammalian MARCKS binds to actin filaments and cross-links them in vitro (Hartwig et al., 1992). The corresponding domain of Xenopus MARCKS is 100% identical (Shi et al., 1997), suggesting that it may also interact with F-actin. At first, we examined the colocalization of MARCKS with F-actin. Cells expressing MARCKS-Venus were dissociated from DMZ explants and cultured on an FN-coated dish. The cells were then fixed and F-actin was stained with phalloidin. As shown in Fig. 3 A, MARCKS and cortical actin were colocalized.
We then constructed two mutants, GA and SD (Fig. 3 B). GA is an unmyristoylated mutant in which the second glycine residue is replaced with alanine. SD is a pseudophosphorylation mutant whose potential phosphorylation sites were replaced with aspartic acid, which is expected not to bind to actin filaments (Hartwig et al., 1992). To detect F-actin, we used the F-actin–binding domain of moesin fused to red fluorescent protein (RFP; Campbell et al., 2002), designated RMA (RFP-moesin actin–binding domain). It has been shown biochemically that this domain binds to F-actin (Turunen et al., 1994; Pestonjamasp et al., 1995). In Drosophila embryos, the corresponding domain of moesin fused with GFP was successfully used to analyze actin dynamics (Dutta et al., 2002). We confirmed that our construct (Venus-moesin actin-binding domain) colocalized with stress fibers and cortical actin stained with phalloidin in CHO cells (Fig. 3 C).
In Xenopus embryonic cells, the RMA was localized to the cell cortices, and cytochalasin B treatment, which disrupts actin filaments, dispersed the RMA to the cytoplasm (Fig. 3 B). This indicates that the RMA should be useful for monitoring F-actin dynamics. When these MARCKS-Venus genes were expressed, the wild-type and GA forms were associated with the plasma membrane and colocalized with RMA, but SD was in the cytoplasm. When the cells were treated with cytochalasin B, GA dispersed to the cytosol with RMA, whereas wild-type remained on the membrane. This result suggested that the association of MARCKS with the membrane was regulated by two mechanisms, myristoylation and binding to the cortical actin. We also found that GA and SD mutants did not inhibit gastrulation movement when they were overexpressed (Fig. 3 B). These mutants also did not rescue the embryo phenotype caused by MARCKS Mo. These results suggest that both myristoylation and actin biding are required for its function.
To test whether MARCKS regulates cortical actin formation, MARCKS Mo was injected into one blastomere near the animal pole of two-cell embryos. Animal caps were isolated at the late blastula stage, fixed, and stained with phalloidin. Membrane-binding RFP was coinjected with Mo for tracing the injected cells. As shown in Fig. 3 D, MARCKS Mo significantly reduced the amount of cortical actin stained by phalloidin. The amount of actin protein was not affected, however, judging from Western blotting and immunocytochemistry using an antipan actin antibody (Fig. 3, E and F). This result suggests that MARCKS plays an important role in cortical actin formation.
The noncanonical Wnt pathway has been implicated in convergent extension. It has been demonstrated that Xdsh, an essential cytoplasmic component in this pathway, regulates cell polarity and protrusive activity in DMZ cells (Wallingford et al., 2000). The Wnt pathway activates RhoA and Rac (Habas et al., 2001, 2003), which have been shown to regulate the protrusive activity (Tahinci and Symes, 2003). Thus, the pathway may directly regulate actin cytoskeletal dynamics. To investigate the relationship between the Wnt pathway and cortical actin, we examined the localization of Xdsh. Cells were dissociated from the DMZ explants expressing Xdsh-Venus, cultured on an FN-coated dish, and stained with phalloidin. As shown in Fig. 4 A, Xdsh was colocalized with cortical actin, even in the lamellipodial and filopodial protrusions.
When RMA was expressed during convergent extension, it was located at the tips of elongated mesodermal cells (Fig. 4 B). This indicates that F-actin is enriched in this region. We showed previously that Xdsh-Venus was also accumulated in the same region (Kinoshita et al., 2003). Mammalian Dishevelled interacts with actin filament through the NH2-terminal DIX domain (Capelluto et al., 2002). To test whether the tip localization of Xdsh was due to the interaction between the DIX domain and F-actin, we tested the localization of Xdsh lacking the DIX-domain (XdshΔDIX). As shown in Fig. 4 B, the XdshΔDIX was located at the tip, indicating that this localization is not due to interaction between F-actin and the DIX domain. This result is consistent with the finding that DishevelledΔDIX can mediate the noncanonical Wnt signaling in Xenopus and zebrafish (Heisenberg et al., 2000; Tada and Smith, 2000). The actin depolymerizing reagent, Latrunculin A, dispersed both RMA and XdshΔDIX to the cytosol. Essentially, the same result was also obtained using cytochalasin B (unpublished data). These results strongly suggest that Xdsh interacts with F-actin either directly or indirectly and mediates the Wnt signaling to the actin cytoskeleton.
To examine whether the Wnt pathway regulates the protrusive activity, we coexpressed Xwnt11 and Xfz7 (Xenopus frizzled-7) in animal cap explants with the membrane-binding RFP. As shown in Fig. 4 C, the coexpression of Xwnt11 and Xfz7 dramatically promoted lamellipodia- and filopodia-like protrusions and it was inhibited by Xdd1, a dominant negative Xdsh mutant (Sokol, 1996; Wallingford et al., 2000). When MARCKS Mo was coinjected, this activity was severely inhibited (Fig. 4 C). It was rescued by coinjection with MARCKS mRNA. In addition, dorsal mesodermal cells expressing dominant-negative Xwnt-11 (Tada and Smith, 2000) significantly reduced the number of protrusions (Fig. 4 D), which is consistent with the observation by Wallingford et al. (2000) that cells expressing Xdd1 maintain significantly fewer stable protrusions. These results strongly suggest that the Wnt signaling pathway regulates cortical actin dynamics and that MARCKS is required for this process.
Here, we have shown that MARCKS plays an essential role in regulating cortical actin dynamics in Xenopus development. MARCKS Mo inhibited cell movements, cell shape change, cell adhesion, and interaction with FN probably through the defect it caused in the cortical actin dynamics. MARCKS is required not only for gastrulation but also for the neural tube formation. When MARCKS Mo was injected into the dorso-anterior blastomeres of eight-cell embryos to target the neuroectoderm, neural tube closure was severely impaired (Fig. S2). It has been shown that MARCKS-deficient mouse shows neural tube closure defect (Stumpo et al., 1995), suggesting the conserved function between frogs and mice. MARCKS may be required for highly organized actin dynamics to effect dynamic tissue reorganization.
The regulation of the cortical actin cytoskeleton by MARCKS may be important for a proper cellular response to signals such as Wnt and the FN/integrin pathways. It is also possible that these signaling pathways regulate the activity of MARCKS. It has been shown that MARCKS is a PKC substrate. PKC has been involved in the noncanonical Wnt pathway (Sheldahl et al., 1999) and the integrin pathway (Vuori and Ruoslahti, 1993). It would be interesting to determine how the activity of MARCKS is regulated during development.
Materials And Methods
Plasmids, RNA synthesis, and Morpholino oligos
Procedures for the plasmid construction, RNA synthesis and sequences of Morpholino oligos were described in the online supplemental material. The RFP plasmid is a gift from R. Tsien (University of California, San Diego, CA).
In situ hybridization and RT-PCR analysis
In situ hybridization in Xenopus was performed as described by Harland (1991). For RT-PCR analyses, RNA from the explants was prepared with TRIzol (Life Technologies). cDNA was synthesized with reverse transcriptase (TRT-101; Toyobo). Sequences of the primers were described in the online supplemental material.
Whole-mount immunostaining and Western blotting
The procedure for whole-mount immunostaining was performed as described in Kurata et al. (2001). The antibodies were MZ15 for notochord (a gift from F. Watt, Imperial Cancer Research Fund, London, UK) and 12/101 for somites (Development Studies Hybridoma Bank). Western blotting was performed using a mouse monoclonal antipan-actin antibody was purchased from NeoMarkers (MS-1295-P0).
Dissecting explants and cytological observations
For the animal cap explants, MARCKS mRNA or Mo was coinjected with 0.5 pg activin mRNA into the animal pole of two-cell embryos. The animal cap was dissected from stage-9 embryos. For DMZ explants, mRNA or a Mo was injected into the two dorsal blastomeres of four-cell embryos. Explants were isolated at stage 10+. These explants were cultured in 1× Steinberg's solution until sibling embryos reached stage 17. To dissociate cells from the explants, the explants were incubated in the Ca2+-Mg2+–free medium for 2 h. For the cytological observation, explants and dissociated cells were cultured in 1× Steinberg's solution on an FN-coated dish (4000–030; Iwakil), or on a cover glass coated with FN (∼1 μg/cm2, F1141; Sigma-Aldrich). To stain F-actin, cells were fixed in 4% PFA and stained with PBS 0.5% Triton X-100 containing a 40-fold dilution of BODIPY 581/589 phalloidin (B-3416; Molecular Probes) or Alexa Fluor 488 phalloidin (A-12379; Molecular Probes). For confocal microscopy, images were captured using 510 software (Carl Zeiss MicroImaging, Inc.). All images were prepared for publication using Adobe Photoshop software.
Online supplemental material
Fig. S1 shows that MARCKS Mo specifically inhibits MARCKS protein synthesis. Fig. S2 shows that MARCKS Mo inhibits neural tube closure. Supplemental material is available online.
We thank Dr. Toshihiko Fujimori for valuable discussion and information.
This work was supported by grants from the Ministry of Education, Science and Culture of Japan to N. Ueno and N. Kinoshita.
The online version of this paper includes supplemental material.
Abbreviations used in this paper: DMZ, dorsal marginal zone; FN, fibronectin; MARCKS, myristoylated alanine-rich C kinase substrate, mb-Venus, membrane-binding Venus; Mo, Morpholino oligonucleotide; RFP, red fluorescent protein; RMA, RFP-moesin actin; Xdsh, Xenopus Dishevelled; XMLP, MARCKS-like protein.