Mammals have skin; insects have cuticle — both function as barriers. It goes without saying that simple wounds in these integuments must be fixed, and William McGinnis' group have defined an evolutionarily conserved pathway, involving the Grainy head (Grh) transcription factor and extracellular signal-regulated kinase (ERK), through which this is achieved.

The authors studied two genes — Ddc and ple — in Drosophila melanogaster. Ddc encodes dopa decarboxylase and ple encodes tyrosine hydroxylase; both proteins are involved in generating the integument of insects. After wounding late D. melanogaster embryos, Ddc and ple transcripts began accumulating in epidermal cells near the wound within 30 minutes, indicating that they were directly targeted by wound-induced signalling. A sequence from −1.4 kb to the transcriptional start of Ddc was required for this response. A tagged reporter construct containing this sequence was activated in a decreasing graded response surrounding the wound, indicating that a signal produced at the injury site probably activated the wound response enhancer dose dependently.

This wound response enhancer didn't require the activity of signalling pathways involved in innate immunity (which protect against infections through wounding), nor did it require the zygotic functions of Jun N-terminal kinase (JNK), Jun or Fos, which are involved in dorsal closure during embryogenesis. What, then, binds to the enhancer region?

One protein isoform encoded by grh — Grh-N — is expressed in barrier epithelia, and zygotic mutations in grh result in a phenotype similar to Ddc and ple mutants. The authors therefore tested a potential function for Grh-N in activating the −1.4-kb Ddc wound response element. The −1.4-kb Ddc reporter was only weakly activated in aseptically wounded grh mutant embryos, and such wound sites couldn't regenerate normal cuticle.

As well as two evolutionarily conserved sites that bind the Grh transcription factor, the Ddc wound response enhancer contains several other consensus binding sites for other transcription factors, including activator protein-1 (AP-1) and ETS, as well as a GGGGGATT motif. A 3-kb fragment of ple also includes such sites, and strongly activates reporter expression around wounds.

McGinnis and colleagues also noticed an increase in phosphotyrosine staining and ERK activation in cells near aseptic wounds. ERK was activated in sites surrounding embryonic wounds within 30 minutes of injury. ERK was still activated in grh mutants, but inhibiting ERK activation decreased the activation of the Ddc wound response reporter. So ERK might somehow transduce the wound signal to Grh, and the aim is to find out how. Once Grh binds to, and activates, Ddc and ple near wounds, crosslinking molecules that repair the epithelial barrier kick into action. In another study in Science, Ting et al. report that a mouse orthologue of grh, Grainy head-like-3 (Grhl3), is needed to form and maintain the epidermal barrier in mice. One target of Grhl3 is the gene for transglutaminase-1, an enzyme that, like the Ddc and ple gene products, is involved in crosslinking epidermal proteins.