This research program focuses on the role of intercellular communication via peptide growth factors in animal development. Using the tools of genetics and molecular biology, our goal is to describe biochemical pathways through which growth factors act to modulate morphology, gene expression and cell behavior. The major focus is on the Transforming Growth Factor beta family, with lesser focus on the Wnt and Hedgehog pathways of secreted signaling molecules. Results are relevant to the use of growth factors as cytokines in the manufacture of cellular products as well as therapeutic agents in and of themselves. Furthermore, understanding the stability of differentiated cells and their capacity to be reprogrammed by their environment is a major concern in use of cellular therapies. Our work uses the fruitfly Drosophila melanogaster as a model system. All growth factors signaling pathways examined to date (TGF-beta, Wnt, Hedgehog, FGF, EGF) are highly conserved between Drosophila and mammals. Major projects: a) Use of genetic interaction screens to identify components of the TGF-b signal transduction pathway. The major TGF-beta family member in Drosophila, the product of the decapentaplegic (dpp) gene, is the TGF-beta family member best characterized by mutational analysis. Mutant phenotypes exist that reflect the many requirements for this growth factor during embryogenesis. We study a class of viable mutations that affect the head capsule of the adult fruitfly. The severity of this class of mutations is closely linked to the functional level of dpp signaling. They therefore serve as a barometer of signal transduction pathway activity and have been used to screen the Drosophila genome for loci that interact with dpp in this pathway. One of the screens used an existing bank of mutations created by P element transposon-tagging to facilitate rapid recovery of identified loci, while a second used an ordered collection of genetic deficiencies covering the majority of the Drosophila genome. Using the completed sequence of the Drosophila genome, we have been able to correlate many interacting loci directly with known genes and putative protein coding regions. Some of the interacting mutations are in previously studied genes, such as transcription factors, cell cycle regulators, cellular adhesion molecules, and chromatin remodeling factors. We have begun analysis of one of our best characterized and most strongly interacting genes, the product of the Drosophila odd-paired (opa) locus. Odd-paired encodes a protein homologous to the zinc-finger class of transcription factors. Loss-of-function alleles of opa interact dominantly with the dpp head capsule class of mutations to specifically alter the adult Drosophila head, and we have found that genetic removal of active opa protein using a temperature sensitive mutation also alters adult head formation in a similar manner. No other structures in the adult fly are affected, indicating that opa and dpp specifically act together in the formation of the adult head. In vertebrates, a related group of proteins, the Zic family of transcription factors, is involved in head and brain formation, suggesting conservation of this role in evolution. Our goal is to define the biochemical interaction between intercellular communication via the TGF-beta pathway and alterations in gene expression caused by the opa transcription factor. b) Biology of the head capsule phenotype. The phenotype caused by mutations used for the genetic screens described above is limited to the head capsule of the adult fly. Various external structures such as the eye and sensory organs are reduced, eliminated or duplicated. The adult head derives from paired epithelial sacs called imaginal discs that are elaborated during embryogenesis and then undergo final differentiation at the metamorphosis of the pupa. Imaginal discs have become a popular model system for studying signal transduction as they have complex pattern formation, but unlike the embryo, also undergo growth, allowing this aspect of signal transduction to be studied. We have identified at the DNA sequence level the lesions associated with our head capsule mutations. We are currently analyzing the region associated with these lesions for the presence of cis regulatory DNA by the creation of transgenic flies bearing Beta-galactosidase reporter constructs. Transgenic flies bearing six different constructs have been created and analyzed in the last year. We have found that, as expected, the DNA region altered in head capsule mutations drives gene expression primarily in the eye antennal disks, primordia of the adult head. In addition, the expression pattern is consistent between the structures altered in the mutant phenotype and the established developmental fate maps for the adult head. We have been able to largely delineate the extent of the DNA region responsible for the head capsule phenotype, and interestingly, it is strictly separated from other cis-regulatory regions, unlike the interspersed pattern seen with other regulatory aspects of the dpp gene. We had previously sequenced one specific mutation, hc1, and found that its lesion is a 17 base pair (bp) deletion that removes putative binding sites for several homeodomain transcription factors. This deletion is within an 800 bp area containing many such putative homeodomain protein binding sites. We created a construct bearing only this region of clustered putative binding sites and found it was sufficient to drive Beta-galactosidase expression in the expected area of the imaginal disk, identically to larger constructs containing this region. This strongly suggested that this cluster of homeodomain binding sites represented a critical element in the head capsule cis regulatory region. In addition, we created an equivalent construct containing DNA from the hc1 mutation. This construct does not drive expression of Beta-galactosidase, further supporting the interpretation that the hc1 mutant phenotype is caused by the loss of function of dpp in the eye antennal disk as a result of disruption of critical homeodomain binding sites. We are currently testing this hypothesis by singly and multiply altering putative homeodomain binding sites within a specific Beta-galactosidase reporter construct to demonstrate that these sites are actually necessary and sufficient to produce gene expression from this construct. Auxiliary to these analyses, we have examined the expression of our Beta-galactosidase reporter constructs in genetic backgrounds that have either lost expression of specific candidate homeodomain genes in the head, or in which these genes are ectopically expressed. We have not yet identified a homeodomain transcription factor that can modify expression of our reporter constructs; however, these analyses are still on going. Our genetic analyses and the expression patterns of our Beta-galactosidase reporter constructs suggest that dpp's head capsule phenotype reflects a requirement for dpp to pattern the ventral surface of the head, and perhaps to play a role in aligning and fusing the two epithelial sheets of the paired eye/antennal disks when the adult head is formed from them at metamorphosis. This is an area of fly development that is very poorly understood. Most of the genetic interactions that have been described for the elaboration of fly tissues are conserved across species line to vertebrates, thus it is likely that data derived from our analysis of Drosophila head development will shed light on genetic interactions relevant to human head (and brain) development. In addition, the fusion of imaginal disks at metamorphosis is considered to be a model for the interaction of such epithelial sheets in the construction and regeneration of tissues, particularly wound healing. c) Transcriptional control of a TGF-beta responsive gene. Dpp participates in the formation of the alimentary system of the fly larva, particularly in the morphogenesis of the midgut. Over the past 10 years I have described the factors responsible for dpp's expression in this tissue. The enhancer elements that control transcription are directly controlled by the HOX ortholog Ultrabithorax, a homeodomain-containing transcription factor. We have also shown that the WNT pathway controls transcriptional repression though the HMG box factor Drosophila TCF. Our current focus is on the homeodomain proteins extradenticle and homothorax, which are the fly homologs of the Pbx and Meis oncogenes in humans. These DNA binding proteins are thought to act as essential cofactors for the action of HOX genes. Our data, however, indicate that they have roles independent of HOX genes, and are capable of regulating gene expression independently. We have completed an analysis that shows that the spatially specific domains of expression of the dpp gene found in the developing midgut are the result of direct activation and indirect inactivation by extradenticle and homothorax.
