The NC plays a critical role in the developmental of the vertebrate head, face and jaws, providing the bulk of the craniofacial skeleton as well as peripheral nervous system and other cranial tissues. Normal craniofacial development depends on proper induction, migration and differentiation of NC cells and derivatives. Deficiencies at any of these steps, whether due to intrinsic defects in NC itself, or in failure of NC cells to interact properly with adjacent tissues, can lead to birth defects: up to a third of all congenital malformations are craniofacial in nature. In this project we have used a key NC control factor, TFAP2, as a molecular tool to identify target genes encoding proteins that are required for NC migration and differentiation into craniofacial tissue. The three genes that comprise the focus of this research are either themselves novel (PCNS and Inca) or have not previously been shown to function in NC (MyoX), and have an excellent potential to yield new insights into craniofacial development. In particular, our findings with Inca promise to lead to a new link between cytoskeletal regulation through small GTPases and PAK4 and regulation of gene expression through MAPK signaling. In the previous year we completed initial studies on PCNS and showed this molecule to be essential for migration of NC cells out of the neural folds. Work on this project has been temporarily suspended in our lab, although other groups are pursuing different aspects of PCNS function. Our focus has been on the other two genes, Inca and MyoX.? ? Inca: Using yeast two-hybrid screening we discovered that Inca interacts with p21-activated kinase 4 (PAK4). This kinase is part of a family of proteins that mediate cell-cell signaling through the small GTPase molecules Rac, Rho and Cdc42. We have found that stable expression of mouse Inca in NIH3T3 cells results in inhibition of PAK4 phosphorylation. PAK4 has been reported to bind to actin filaments and regulate their dynamic remodeling through LIM kinase 1 and cofilin. If Inca controls PAK4 activity, this would predict an effect on the microfilament network in cells. We have tested and confirmed this hypothesis by examining phalloidin staining of a stable Inca expressing NIH3T3 line and showing that these cells have much more extensive stress fiber networks than an empty vector control line. Furthermore, this phenotype could be reversed by expression of a constitutive activated PAK4. An additional test of this model is based on reports that PAK4 phosphorylates two regulatory serine residues (S67 and S810) on GEF-H1S, a regulator of Rac and RhoA activation. This leads to the displacement of Rho in favor of Rac, reducing the level of GTP-Rho and increasing GTP-Rac, favoring formation of lammelipodia over stress fibers. This suggests the possibility that Inca might be acting to inhibit PAK4 kinase activity, driving the Rac/Rho activation ratio to favor stress fibers. We confirmed this hypothesis by using a standard Rho activation pulldown assay on control and Inca expressing cell lines, revealing increased RhoGTP levels in the latter. Based on these findings we conclude that Inca regulates actin polymerization at least in part via PAK4. ? ? Several lines of evidence link the regulation of microfilaments and microtubules, suggesting the possibility that Inca might also affect microtubule dynamics, either indirectly by altering microfilaments or by a more direct mechanism. This was also investigated in the Inca expressing cell lines. While we found little evidence for changes in microtubule stability, the level of alpha-tubulin acetylation was markedly reduced in the Inca cell lines. This may also be mediated via PAK4, although alternate pathways are possible.? We are interested in translating our findings with cultured mouse cells to the context of the vertebrate embryo. We have previously shown that Inca overexpression, in a synergistic interaction with PAK4, can disrupt the purse string actin cables associated with wound healing in the embryo, and Inca knockdown inhibits migration in NC cells suggesting at least some cytoskeletal functions may be shared. Consistent with this notion, we have found a substantial reduction in stress fibers in the Inca knockdown NC cells. Inca overexpression also inhibits alpha tubulin acetylation in Xenopus ectoderm, and interestingly, strongly enhances the expression in these cells of the phosphorylated form of cofilin, a key regulator of actin polymerization. Preliminary experiments suggest that this is not mediated by PAK4, so another regulatory pathway is likely to be involved.? There is a 35 amino acid domain region roughly in the middle of the Inca protein that is conserved among the Inca genes in vertebrates. We have termed this the Inca Box (IB). To test the importance of this domain in Inca overexpression phenotypes an internal deletion was generated (deltaIB). Injection of this RNA at multiple doses resulted in very slight changes in ectodermal pigment granule distribution, and did not inhibit wound healing, in contrast to the full-length protein, suggesting the IB was essential for the molecular basis for these phenotypes. Interestingly, overexpression of Inca and Inca knockdown both inhibit gastrulation movements, deltaIB has the opposite effect, enhancing elongation of activin-treated ectoderm. This suggests that removing the IB generates a neomorphic molecule rather than constituting a loss-of function. Furthermore, deltaIB is unable to interact with PAK4, so its effects on the embryo are likely due to additional signaling pathways. ? ? One possible pathway is the MAPK cascade: We have found that in many (but not all) experiments, deltaIB RNA strongly activates Xbra expression. This is accompanied by a robust activation of p42/ERK2, a component of MAPK cascade signaling. This result, suggests that in addition to its potential role as a cytoskeletal regulatory factor, Inca may also function to modulate signal transduction and subsequent gene expression associated with mesoderm induction. In the coming year we will determine how deltaIB, and possible also Inca, influence MAPK signaling by using standard inhibitors of cascade components. ? ? MyosinX: Knockdown of zygotic MyoX expression using splice-inhibiting MOs partially inhibited NC migration in vivo. Migration of NC cells in vitro on fibronectin is a more direct assay for migration compared to in situ hybridization to fixed embryos, and also appears to be more sensitive to perturbations. We found a significant inhibition of migration in vitro from MyoX knockdown. The MyoX knockdown NC cells adhere rather poorly to the fibronectin substrate. Since MyoX is known to interact with beta-integrins comprising part of the fibronectin receptor, our hypothesis is that integrin transport may be disrupted in MyoX knockdown NC. To test test for such an interaction in the NC, we co-expressed beta3 integrin labelled with mCherry and MyoX labeled with GFP in frog embryos, then excised NC tissue followed by live confocal imaging after the NC cells had spread on a fibronectin substrate. Co-localization was observed, particularly at the tips of filopodia, and this was disrupted by MyoX knockdown, supporting our hypothesis that MyoX is required for integrin transport and/or activation in migrating NC. This is the first demonstration of such a function for this protein in embryonic cells.
Showing the most recent 10 out of 18 publications
