Developmental biology

SMAD proteins and mammalian anatomy

How does the mammalian body plan, with its different anatomical structures and varied tissues, arise during development? The patterns and anatomical changes that occur have been well described, but the cell-made factors and signalling mechanisms that drive these changes have been more elusive. A paper by Nomura and Li1 on page 786 of this issue and reports in Cell2 and Genes and Development3 describe some of these factors and show that the signal transducers Smad2 and Smad4 are vital in early development.

Members of the transforming growth factor-β (TGF-β) superfamily regulate many important developmental processes, such as mesenchymal differentiation, skeletal morphogenesis and skin formation. The roles of most of these secreted proteins are poorly characterized, but others — such as the activins and several ‘bone morphogenetic proteins’ (BMPs) — are known to regulate specific events in both early and late development. Exactly how these factors act remained unclear until three years ago, when SMADs were discovered as intracellular proteins that mediate the effects of signalling from extracellular TGF-β-related factors (Fig. 1, overleaf). Now Nomura and Li1, Waldrip et al.2 and Sirard et al.3 describe the developmental consequences of targeted inactivation of the genes encoding the Smad2 or Smad4 proteins.

Figure 1: Signalling through SMAD proteins.
figure1

Receptor complexes for transforming growth factor-β (TGF-β) and related factors, such as activins and bone morphogenetic proteins (BMPs), consist of two transmembrane serine/threonine kinases, the type II and type I receptors. Binding of activin, TGF-β or BMPs to the receptor complex results in phosphorylation of type I receptors by type II receptors. The activated type I receptors in turn phosphorylate, and thereby activate, a select set of SMADs: Smads 1 and 5 are activated by BMP receptors, whereas Smads 2 and 3 are phosphorylated by receptors for activin and TGF-β. These SMADs then dissociate from the receptors to form complexes with Smad4/DPC4, which is a putative tumour-suppressor protein. Smad4 is essential for signalling by TGF-β-related factors. These SMAD complexes move into the nucleus, where they may activate, or repress, transcription of different, defined genes14. Smads 6 and 7 (not shown) inhibit signalling, probably by competing for receptor binding and/or for association with Smad4/DPC4 (ref. 14). Phosphorylation is represented by a circled ‘P’.

Gastrulation is an early developmental process in which the three primary layers of the animal body (ectoderm or outer layer, endoderm or inner layer, and mesoderm) are formed and positioned for further development. In mice, gastrulation starts with the formation of the primitive streak — a localized thickening of the epiblast (embryonic ectoderm) occurs and marks what is to be the posterior end of the mouse embryo. As gastrulation proceeds, the primitive streak extends anteriorly and epiblast cells move through the primitive streak to form mesoderm. Expression of both the transcription factor Brachyury and the TGF-β-related protein nodal marks the cells of the primitive streak and the onset of mesoderm formation. The mesodermal cells then develop into several extra-embryonic and embryonic structures (Fig. 2).

Figure 2: Representation of a mouse embryo at early gastrulation (embryonic day of development 6.5), indicating the different cell layers.
figure2

The site of mesoderm formation is shown in purple.

Studies of the phenotypes of Smad2-defective mice1,2 now reveal that Smad2 signalling is essential for embryonic mesoderm formation and the establishment of anterior-posterior polarity. Nomura and Li1 report the absence of mesoderm in Smad2-deficient embryos, and Waldrip et al.2 see extra-embryonic mesoderm without formation of embryonic mesoderm, suggesting a less severe phenotype. These differences in phenotype may be due to differences in genetic background or in the strategy used by the authors to target the Smad2 gene for deletion. Smad2 is also required to establish the anterior-posterior polarity, probably because of its function in restricting the site of primitive-streak formation. So Smad2 signalling is one of the earliest-known events required for the formation of mesoderm and anterior-posterior patterning.

The essential role of SMAD signalling in early development is also shown by the gastrulation defect of Smad4-deficient embryos, reported by Sirard et al.3. Although these embryos also lack mesoderm, they are morphologically different from Smad2-deficient embryos and have highly abnormal visceral endoderm development (the position of the visceral endoderm is shown in Fig. 2). This phenotype may be due to a vital need for Smad4 expression at an earlier stage than Smad2, possibly to allow BMP-induced SMAD signalling during early gastrulation4.

The early defects and lack of mesoderm formation in the absence of Smad2 raise two related questions. First, why is this phenotype more severe than the phenotypes resulting from inactivation of TGF-β-related factors that activate Smad2? There are more than 40 TGF-β-related factors, but only five SMADs (Smads 1, 2, 3, 5 and 9) are activated by the ligand-stimulated receptors. This means that many TGF-β-related factors may signal through the same SMADs. Smad2 is activated by two activins and three TGF-βs, and probably by several other TGF-β-like factors, so inactivation of Smad2 may abolish signalling from a set of TGF-β-related factors.

Second, which TGF-β-related ligand(s) induce Smad2 activation before mesoderm formation? Although activins have been implicated in the formation of mesoderm and activate Smad2, no gross defects in mesoderm formation occur in the absence of activins A and B (ref. 5). In contrast, inactivation of BMP-4 (ref. 4) or nodal6 results in severe early defects and impaired mesoderm formation. BMP-4 signals through Smads 1 and/or 5, and not through Smad2, so is unlikely to be the ligand that activates Smad2. This brings us to a possible connection between nodal and Smad2. Nodal is expressed before formation of the primitive streak, and its inactivation results in defects in this process6. The receptors and SMADs that mediate nodal activity are not known. Embryos lacking Smad2 or nodal have similar defects, so Nomura and Li1 analysed the phenotypes of mice lacking one copy of the Smad2 gene and one copy of the nodal gene. Their findings indicate that nodal may signal through Smad2, although biochemical analyses are required to test this possibility. Other ligands, possibly of maternal origin, are also likely to be required for the Smad2 activation that leads to the formation of mesoderm.

Nomura and Li1 also found that developmental changes depend on the amount of Smad2 activity. A defective phenotype is usually only apparent when both of a pair of genes are inactivated. But the presence of only one functional Smad2 gene, and the resulting lower-than-normal levels of Smad2 protein, also results in developmental defects, albeit much less severe than those seen when both genes are inactivated. This striking dosage-dependence of Smad2 signalling — apparent both in early development and at later stages — is consistent with the observation that, in the African clawed toad Xenopus laevis, different mesodermal markers and cell types are induced by different concentrations of activin7, nodal8 or Smad2 (ref. 9). Different thresholds for transcriptional activation of individual genes by Smad2, and competition of Smad2 with other SMADs for binding to gene promoters, are likely to be the basis for these dosage-dependent changes.

Finally, the new results1 highlight the role of TGF-β-related factors in defining left-right asymmetry. Visceral organs and the position and organization of the heart are normally asymmetric. Little is known about how this left-right patterning occurs, but TGF-β-related factors such as nodal10,11, lefty12 and cNR-1 (ref. 13) are asymmetrically expressed and may regulate left-right patterning. SMAD signalling may relay information from these ligands; consistent with this idea, decreased Smad2 signalling resulted in various abnormalities that reflect disturbances in the left-right asymmetric pattern1.

The inactivation of Smads 2 and 4 has made it possible to study how this class of signalling effectors is involved in establishing the body plan and in driving many developmental changes. The results emphasize the important functions of TGF-β-related factors at many stages in development, from the fertilized egg to the adult organism.

References

  1. 1

    Nomura, M. & Li, E. Nature 393, 786–790 (1998).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Waldrip, W. R.et al. Cell 92, 797–808 (1998).

    Google Scholar 

  3. 3

    Sirard, C.et al. Genes Dev. 12, 107–119 (1998).

    Google Scholar 

  4. 4

    Winnier, G. M., Blessing, M., Labosky, P. A. & Hogan, B. L. M. Genes Dev. 9, 2105–2116 (1995).

    Google Scholar 

  5. 5

    Matzuk, M. M.et al. Nature 374, 354–356 (1995).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Conlon, F. L.et al. Development 120, 1919–1928 (1994).

    Google Scholar 

  7. 7

    Green, J. B., New, H. V. & Smith, J. C. Cell 71, 731–739 (1992).

    Google Scholar 

  8. 8

    Jones, C. M.et al. Development 121, 3651–3662 (1995).

    Google Scholar 

  9. 9

    Baker, J. C. & Harland, R. M. Genes Dev. 10, 1880–1889 (1996).

    Google Scholar 

  10. 10

    Collignon, J., Varlet, I. & Robertson, E. J. Nature 381, 155–158 (1996).

    ADS  CAS  Article  Google Scholar 

  11. 11

    Lowe, L. A.et al. Nature 381, 158–161 (1996).

    ADS  CAS  Article  Google Scholar 

  12. 12

    Meno, C.et al. Genes to Cells 2, 513–524 (1997).

    Google Scholar 

  13. 13

    Levin, M.et al. Cell 82, 803–814 (1995).

    Google Scholar 

  14. 14

    Heldin, C.-H., Miyazono, K. & ten Dijke, P. Nature 390, 465–471 (1997).

    ADS  CAS  Article  Google Scholar 

Download references

Author information

Affiliations

Authors

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Derynck, R. SMAD proteins and mammalian anatomy. Nature 393, 737–739 (1998). https://doi.org/10.1038/31593

Download citation

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.