The Drosophila embryo is patterned by a complex interplay of zygotically expressed genes and maternally supplied components, a large number of which have been identified. However, many maternal components are encoded by essential zygotic genes whose maternal effects are not amenable to conventional genetic analysis. Investigation of such genes requires the generation of homozygous mutant germ cells in chimeric females, and analysis of their embryos. The recent development of techniques which allow the efficient generation of germline clones has made the screening of zygotic lethal mutations for maternal effects more feasible. I have generated a collection of X-linked zygotic lethal mutations and used FLP recombinase catalysed mitotic recombination to look for maternal effects affecting segmentation. Two mutations have been recovered which have maternal effect phenotypes similar to those of the pair-rule segmentation genes. The leprechaun mutation affects oogenesis, so fertile females are very rare, preventing straightforward phenotypic analysis of the segmentation phenotype. Attempts to generate a rescuing duplication are described. The second mutation, stunted (sun), initially gave rise to a segmentation cuticle phenotype. Subsequent attempts to reproduce the phenotype were unsuccessful as it was masked by a severe reduction in the amount of cuticle secreted, a phenotype characteristic of the neurogenic genes. Detailed analysis revealed that the primary lesion affects neither segmentation or neurogenesis. Rather, sun⁺ is required for cellularisation of the syncytial blastoderm and for the localisation of actin to 'caps' above the syncytial nuclei. Cloning of a candidate gene for stunted revealed a predicted protein product limited homology to cyclins. In addition to searching for novel segmentation genes, potential proteinprotein interactions of the segmentation gene hairy's protein product were also investigated, and a model is presented for its mode of action as a transcriptional repressor.

The blastoderm embryo of Drosophila melanogaster consists of a unicellular syncytium with a large number of peripheral nuclei. The cytoplasm surrounding each peripheral nucleus is compartmentalised into apical periplasm above each nucleus and basal periplasm below it. The expression of different genes in the syncytial blastoderm is crucial for the genetic control of development. The pair-rule genes are involved in controlling the pattern of metamerisation of the embryo. Pair-rule mRNAs are expressed in alternate metameres, in a pattern of stripes. Within each stripe, mRNA is found in the apical periplasm of the syncytial blastoderm. By analysing the distribution of mRNA of a number of hybrid constructs, I show that the 3' untranslated part of three pair-rule genes are required for the apical localisation of their transcripts. A 1.2kb region in the 3' end of fushi tarazu (ftz), a 700bp region in the 3' end of hairy (h) and a 160bp fragment of the 3' untranslated part of the even-skipped (eve) pair-rule gene are shown to contain apical localisation signals. I show that the mechanism of apical localisation is unlikely to involve a cytoplasmic process and that the 3' untranslated part of the bicoid (bed) gene contains sequences necessary for apical localisation. I propose that apical localisation involves a nuclear mechanism which exports mRNA from the apical side of the nuclear membrane. I demonstrate that apical localisation is achieved by an RNA-mediated process and not by a DNA-mediated mechanism. Finally, I demonstrate that the intracellular localisation of transcripts encoding cytoplasmic proteins influences the distribution of the protein in the periplasm. I propose that the function of apical localisation is to limit the diffusion of pair-rule proteins so that the pattern of protein expression resembles precisely the transcriptional domain.

No description available.