The eye disc is specified to form an eye by a hierarchy of transcription factors headed by two Pax-6 genes. Subsequently, signaling by the secreted Hedgehog (Hh) and Decapentaplegic (Dpp) proteins is required for photoreceptor differentiation to initiate at the posterior margin of the disc and progress anteriorly. My lab has used a genetic mosaic technique to screen for new mutations affecting the ability of cells in the eye disc to differentiate as photoreceptors. We have screened 80% of the genome and are now in the process of characterizing the novel genes. Several of the genes we have identified appear to influence Hh function. One such novel gene, act up (acu), encodes a homologue of yeast cyclase-associated protein. acu is required to prevent the polymerization of actin monomers, and we found that acu mutations also lead to premature photoreceptor differentiation. The normal apical constriction of cells in the morphogenetic furrow does not occur in acumutants; our results suggest that this constriction is required to limit the spread of Hh and thus prevent cells from differentiating precociously. Another of the genes we identified, sightless (sit), is required for Hh signaling; Hh target gene expression is lost in sit mutants, although the expression of hh itself is unaffected. We showed that sit is required in the Hh-producing cells and its protein product has homology to a family of transmembrane acyltransferases. Since we have also shown that a cysteine residue that is palmitoylated in vertebrate Sonic hedgehog is essential for the function of DrosophilaHh, this suggests that Sit may be the enzyme that palmitoylates Hh. We are now investigating other potential substrates of Sit based on its mutant phenotypes. We also isolated mutations in the hyperplastic discs (hyd) gene, which encodes a HECT domain ubiquitin protein ligase. hydmutations lead to premature photoreceptor differentiation as well as non-autonomous overgrowth of cells in the eye disc. Loss of hydcauses ectopic hhexpression as well as autonomous activation of the Hh pathway. We are investigating the mechanisms by which this occurs. Another gene encodes the myosin binding subunit of myosin light chain phosphatase. We have shown that this enzyme acts by regulating myosin activity and is required to maintain photoreceptors within the eye disc epithelium. Myosin phosphatase activity may be required to prevent neuronal cell bodies from migrating towards their axonal growth cones. We are also following up at least five additional genes from this screen that have phenotypes suggesting interesting patterning defects.

Two of the genes that we identified in the above screen, skuld (skd) and kohtalo (kto), encode Drosophila homologues of the mediator complex subunits TRAP240 and TRAP230. Although this complex appears to be a general transcriptional coactivator and loss of any of its core subunits is cell-lethal, mutations in skd and kto cause specific patterning defects rather than affecting cell growth. Both skd and kto are required for the expression of a subset of the target genes of the Hh, Wingless (Wg) and Notch pathways. They appear to act directly on the target promoters, because target gene expression is not rescued by activating the pathway at the transcriptional level. In addition, they are required to establish the cell affinity differences that allow stable compartment boundaries to form in the wing disc. The two genes have identical phenotypes and their protein products physically interact. This suggests that TRAP240 and TRAP230 constitute a submodule of the mediator complex that functions with specific transcription factors to activate a subset of their target genes. We are now trying to determine the basis for their specificity at the level of both interacting transcription factors and enhancer sequences. In a related project, we have shown that the osagene encodes an ARID-domain-containing protein that binds DNA without sequence specificity and associates with the Brahma chromatin remodeling complex. osa mutations lead to embryonic and wing disc phenotypes that resemble ectopic activation of the Wingless (Wg) pathway; we have shown that osa is required for the repression of several genes regulated by the Wg pathway. We believe that it acts directly on these genes because it functions downstream of the up-regulation of Armadillo and because its action depends on the presence of dTCF binding sites in the target promoter. Other components of the Brahma complex also appear to be required to repress Wg target genes. These results imply that an important function of Wg signaling is to overcome a repressive chromatin structure on target promoters, consistent with its use of an architectural transcription factor, dTCF. Our current goal is to understand the mechanism of Osa function by searching biochemically for proteins that interact with a second highly conserved domain of Osa, and genetically for genes that can modify the effects of Osa overexpression.

The projection of Drosophila photoreceptor axons to the visual areas of the brain forms a topographic map reflecting the visual field. The projection is also cell-type specific; two subgroups of photoreceptors project to different brain areas. To understand the molecular mechanisms guiding these axons, we carried out a genetic screen for mutations that cause photoreceptor axons to project abnormally. We identified mutations in two genes required for photoreceptor R7 to terminate in a deeper layer than R8. We have shown that one of these genes is LAR, which encodes a receptor tyrosine protein phosphatase. LARmutations have a fully penetrant and specific effect on R7 targeting; they do not affect R7 development or the targeting of R1-6 or R8. LAR can function in either R7 or R8 to rescue the R7 phenotype; its intracellular domain is dispensable for its function in R8, suggesting that in this situation it can act as a ligand. The second gene, out of step (oos), encodes an a-liprin, vertebrate homologues of which cluster the LAR receptor. oos and LAR have identical phenotypes, suggesting that clustering could be essential for LAR function or that liprin may have signaling activity. We are testing both possible functions for liprin.