Integration of Smad and MAPK pathways: a link and a linker revisited

Abstract

Cells develop by reading mixed signals. Nowhere is this clearer than in the highly dynamic processes that propel embryogenesis, when critical cell-fate decisions are made swiftly in response to well-orchestrated growthfactor combinations. Learning how diverse signaling pathways are integrated is therefore essential for understanding physiology. This requires the identification, in tangible molecular terms, of key nodes for pathway integration that operate in vivo. A report in this issue, on the integration of Smad and Ras/MAPK pathways during neural induction (Pera et al. 2003), provides timely insights into the relevance of one such node. Pera et al. (2003) report that FGF8 and IGF2—two growth factors that activate the Ras/MAPK pathway— favor neural differentiation and mesoderm dorsalization in Xenopus by inhibiting BMP (Bone Morphogenetic Protein) signaling. Mesoderm is formed from ectoderm in response to Nodal-related signals from the endoderm at the blastula stage and beyond (Fig. 1; for review, see De Robertis et al. 2000). BMP induces differentiation of ectoderm into epidermal cell fates at the expense of neural fates, and it ventralizes the mesoderm at the expense of dorsal fates (for review, see Weinstein and HemmatiBrivanlou 1999; De Robertis et al. 2000). Accordingly, neural differentiation and dorsal mesoderm formation are favored when BMP signaling is attenuated. Noggin, Chordin, Cerberus, and Follistatin, secreted by the Spemann organizer on the dorsal side at the gastrula stage, facilitate the formation of neural tissue by sequestering BMP (Weinstein and Hemmati-Brivanlou 1999; De Robertis et al. 2000). Experimentally blocking BMP signaling with a dominant-negative BMP receptor has a similar effect of promoting ectoderm neuralization (Weinstein and Hemmati-Brivanlou 1999). As it turns out, neural induction can also be achieved with FGF (fibroblast growth factor; Kengaku and Okamoto 1993; Lamb and Harland 1995; Hongo et al. 1999; Hardcastle et al. 2000; Streit et al. 2000; Wilson et al. 2000) and IGF (insulin-like growth factor; Pera et al. 2001; Richard-Parpaillon et al. 2002). Injection of transcripts encoding FGF8 or IFG2 into one animal-pole blastomere of a fourto eight-cell embryo results in an expanded neural plate at the injected side (Pera et al. 2003). Surprisingly, expression of a dominant-negative FGF receptor prevents neuralization of ectoderm explants by the BMP blocker Noggin (Launay et al. 1996). Likewise, the potent neuralizing effect of Chordin can be blocked by a dominant-negative FGF receptor or a morpholino oligonucleotide targeting the IGF receptor (Pera et al. 2003). Thus, the neuralizing effect of BMP inhibitors is somehow tied to FGF and IFG signaling. The question is, how? Because FGF8 and IFG2 activate MAPK, Pera et al. (2003) took heed from previous work showing that MAPK inhibits the BMP signal-transduction factor Smad1 (Kretzschmar et al. 1997a). Smad1 is directly phosphorylated by the BMP receptor, resulting in Smad1 activation (Kretzschmar et al. 1997b), and by MAPK in response to EGF, resulting in Smad1 inhibition (Kretzschmar et al. 1997a; Fig. 2). Smad transcription factors mediate gene responses to the entire TGF (Transforming Growth Factor) family, to which the BMPs belong (for review, see Massague 2000; Derynck and Zhang 2003). Smads 1, 5, and 8 act primarily downstream of BMP receptors and Smads 2 and 3 downstream of TGF , Activin and Nodal receptors. Smad proteins have two conserved globular domains—the MH1 and MH2 domains (Fig. 2). The MH1 domain is involved in DNA binding and the MH2 domain in binding to cytoplasmic retention factors, activated receptors, nucleoporins in the nuclear pore, and DNA-binding cofactors, coactivators, and corepressors in the nucleus (for review, see Shi and Massague 2003). Receptor-mediated phosphorylation occurs at the carboxy-terminal sequence SXS. This enables the nuclear accumulation of Smads and their association with the shared partner Smad4 to form transcriptional complexes that are interpreted by the cell as a function of the context (Massague 2000). Between the MH1 and MH2 domains lies a linker region of variable sequence and length. Attention was drawn to this region when it was found that EGF (epidermal growth factor), a classical activator of the Ras/ MAPK pathway, causes phosphorylation of the Smad1 linker at four MAPK sites (PXSP sequences; Kretzschmar et al. 1997a). This prevents the nuclear localization of Smad1 and inhibits BMP signaling. Mutation of these E-MAIL j-massague@ski.mskcc.org; FAX (212) 717-3298. Article and publication are at http://www.genesdev.org/cgi/doi/10.1101/ gad.1167003.