Signal transduction

An eye on organ development

Studies in flies and mice have revealed a surprising way in which cells regulate gene activity, with consequences for our understanding of organ formation during development.

The processes by which cells receive outside signals and transmit them into the nucleus have been so well scrutinized that one might think few surprises remain. But three papers in this issue1,2,3 show otherwise. During development, many of the signalling pathways that regulate organ formation are initiated by growth factors that bind to receptors on the cell surface. This triggers a cascade of events inside the cell that frequently involves the addition or removal of phosphate groups (phosphorylation or dephosphorylation, respectively) from cellular proteins. These proteins include those that regulate gene expression, such as transcription factors and their associated 'co-activators' and 'co-repressors'. Rayapureddi et al.1, Tootle et al.2 and Li et al.3 now report the intriguing finding that one type of transcription factor can itself dephosphorylate other proteins — and that this function is crucial in various examples of organ formation. These results reveal a new way in which cells can fine-tune gene expression.

Among the most remarkable discoveries of the past decade has been the identification of single genes that can induce the formation of entire organs or tissues. In fruitflies (Drosophila melanogaster), for example, misexpression of a single gene called Eyeless, one of several 'retinal-determination' genes that encode transcription factors4, is enough to evoke eyes where eyes shouldn't be ('ectopic' eyes). Eyeless is a strong inducer of ectopic eyes and is required for the expression of other retinal-determination genes, including Eyes absent (Eya), Sine oculis and Dachshund. Eya and Dachshund can each induce ectopic eye tissue, although only weakly. Expressing Eya together with Sine oculis or Dachshund markedly enhances ectopic eye development, and these genes appear to function together in a molecular network5,6. Consistent with this notion, inactivating any of these genes results in flies lacking eyes4. The components of this network have been highly conserved during evolution, with related genes in mammals regulating the development of multiple organ systems. Mutations of the human counterparts cause a variety of congenital disorders7.

What do we know about how the products of these genes interact at the molecular level? First, Sine oculis and its mammalian relatives are DNA-binding proteins that, depending on the context, can either activate8,9 or repress10 transcription. Second, Eya proteins contain highly conserved structural regions known as Eya domains, which mediate interactions with Sine oculis and Dachshund, as well as poorly conserved regions that can activate transcription7 (Fig. 1a). Third, when Eya and Sine oculis are expressed together, they potently activate transcription — even under conditions in which Sine oculis alone acts as a repressor8,9,10. Fourth, Dachshund can activate transcription in yeast6, but it is also related to two transcriptional repressors and can bind co-repressors10.

Figure 1: The eyes (absent) have it.

a, Proteins of the Eyes absent (Eya) family are composed of divergent transcription-activating domains and conserved Eya domains. The Eya domain mediates interactions with the Sine oculis and Dachshund proteins. The new papers1,2,3 show that the Eya domain also works as a phosphatase, removing phosphate groups ('P') from tyrosine amino acids (and possibly3 serine and threonine) in unknown target proteins (and possibly2 from itself). b, c, Effects of Eya on transcription. b, Top, Dachshund acts as a transcriptional repressor, which 'recruits' co-repressors such as NCoR, Sin and histone deacetylases. Bottom, Eya's phosphatase activity converts Dachshund into an activator. The activation complex includes CBP and RNA polymerase II, but the substrate of Eya's phosphatase activity is unknown. c, Eya and Sine oculis together can also activate transcription in a Dachshund-independent fashion. This probably does not require phosphatase activity.

Now, Rayapureddi et al.1, Tootle et al.2 and Li et al.3 reveal a surprising new function for Eya proteins. All three groups started by noting that there is a similarity in amino-acid sequence between the Eya domains and sequence 'signatures' in enzymes of the haloacid dehalogenase (HAD) superfamily. This large family includes a class of dephosphorylating enzymes (phosphatases)11, some of which remove phosphate groups specifically from serine amino acids in target proteins. Notably, HAD proteins carry out catalysis differently from other well-known phosphatases12,13.

The HAD-related signature in Eya domains hinted that the Eya proteins might also be phosphatases. Indeed, all three groups show that the Eya domains dephosphorylate various artificial substrates. Moreover, mutations in key amino acids within Eya domains — including the likely catalytic amino acid, an aspartate — decrease or eliminate catalytic activity. Tootle et al.2 go on to look at several mutant Eya proteins whose phosphatase activity is predicted to be impaired to varying degrees. They find that these proteins are defective in inducing ectopic eyes in flies, and in restoring normal eyes to flies with inactive Eya. Rayapureddi et al.1 confirm that mutating the catalytic aspartate of Eya results in a decreased (though not entirely missing) ability to restore eyes to flies that lack functional Eya.

Li et al.3 provide exciting hints as to why phosphatase activity is important for Eya function. They generate mice that lack Six1, a mammalian relative of Sine oculis, and find that these animals have muscle, kidney and skeletal abnormalities, resulting from defective cell proliferation. They go on to show that the genes encoding c-Myc and glial-derived neurotrophic factor are normally repressed by Six1. Moreover, this repression is potentiated by Dach1 (a mouse relative of Dachshund), and reversed by Eya3. Surprisingly, inactivating Dach1 by various means also impairs Six1-mediated transcription of its target genes. Furthermore, the ability of Eya3 to reverse repression is destroyed by mutating the catalytic aspartate. Li and colleagues' conclusion is as unavoidable as it is remarkable: Dach1 acts as both a co-repressor and a co-activator of Six1 — and Eya proteins, using their embedded phosphatase activities, catalyse the switch from repression to activation (Fig. 1b).

Although there is ample precedent for phosphorylation affecting transcriptional activity, Eya proteins represent the first transcription factors with intrinsic phosphatase activity that are capable of modulating transcriptional complexes. Several key questions remain: notably, which proteins does Eya dephosphorylate, and how is its activity regulated? In this regard, the reported catalytic properties of the various Eya domains studied by the three groups differ dramatically. For example, the catalytic efficiency of mouse Eya3 varies by more than three orders of magnitude1,2,3. Moreover, Rayapureddi et al. and Tootle et al. maintain that Eya3 removes phosphates only from tyrosine amino acids, whereas Li et al. claim that it has dual specificity.

What are these target proteins? Tootle et al. suggest that Eya3 can remove phosphate groups from itself. Li et al. found that Eya3 can dephosphorylate RNA polymerase II — a key transcription enzyme. Further experiments are needed to determine the physiological relevance of both observations. Given that Eya3's phosphatase activity is required to switch Dach1 from a repressor to an activator, another attractive possibility is that Dachshund (or possibly Sine oculis) is an Eya target.

Other differences between the groups' findings may suggest modes of regulation. For example, Tootle et al. measured the activity of a synthetic protein containing an Eya domain that is potentially capable of forming a dimer. This protein was much less active than the isolated Eya domain studied by Rayapureddi and colleagues. Dimerization may be of regulatory significance, because Eya can self-associate9. As Li et al. report, full-length Eya was also considerably less active than the isolated Eya domain, hinting that other domains in the complete protein regulate the Eya domain.

Finally, what is the exact relationship between the catalytic and biological activities of Eya? Tootle et al. looked at Eya proteins that had different mutations in the catalytic domain, and found both decreased phosphatase activity and a diminished ability to generate eyes. But there is no clear connection between the extent of impairment of the two activities. This may reflect details of the experimental design, or, more probably, that Eya has phosphatase-dependent and -independent biological activities (Fig. 1b, c). Alternatively, if an Eya mutant no longer has catalytic activity, but can still associate with its target proteins, then, by sequestering those targets away from potential interacting proteins, it might exhibit some similar effects to the normal phosphatase. Analogous mutants of other phosphatases have such effects13. If so, such Eya mutants might provide a rapid means of identifying physiologically relevant targets of these intriguing new phosphatases.


  1. 1

    Rayapureddi, J. P. et al. Nature 426, 295–298 (2003).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Tootle, T. L. et al. Nature 426, 299–302 (2003).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Li, X. et al. Nature 426, 247–254 (2003).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Treisman, J. E. BioEssays 21, 843–850 (1999).

    CAS  Article  Google Scholar 

  5. 5

    Pignoni, F. et al. Cell 91, 881–891 (1997).

    CAS  Article  Google Scholar 

  6. 6

    Chen, R., Amoui, M., Zhang, Z. & Mardon, G. Cell 91, 883–901 (1997).

    Google Scholar 

  7. 7

    Wawersik, S. & Maas, R. L. Hum. Mol. Genet. 9, 917–925 (2000).

    CAS  Article  Google Scholar 

  8. 8

    Ohto, H. et al. Mol. Cell. Biol. 19, 6815–6824 (1999).

    CAS  Article  Google Scholar 

  9. 9

    Silver, S., Davies, E. L., Doyon, L. & Rebay, I. Mol. Cell. Biol. 23, 5989–5999 (2003).

    CAS  Article  Google Scholar 

  10. 10

    Li, X., Perissi, V., Liu, F., Rose, D. W. & Rosenfeld, M. G. Science 297, 1180–1183 (2002).

    ADS  CAS  PubMed  Google Scholar 

  11. 11

    Aravind, L., Galperin, M. Y. & Koonin, E. G. Trends Biochem. Sci. 23, 127–129 (1998).

    CAS  Article  Google Scholar 

  12. 12

    Barford, D., Das, A. K. & Egloff, M. P. Annu. Rev. Biophys. Biomol. Struct. 27, 133–164 (1998).

    CAS  Article  Google Scholar 

  13. 13

    Tonks, N. K. & Neel, B. G. Curr. Opin. Cell Biol. 13, 182–195 (2001).

    CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Epstein, J., Neel, B. An eye on organ development. Nature 426, 238–239 (2003).

Download citation

Further reading


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.


Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing