The forked (f) gene of Drosophila melanogaster encodes six different transcripts 6.4, 5.6, 5.4, 2.5, 1.9, and 1.1 kb long. These transcripts arise by the use of alternative promoters. A polyclonal antibody raised against a domain common to all of the forked-encoded products has been used to identify forked proteins on two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels and in Drosophila pupal tissues. The antibody stains fiber bundles present in bristle cells for about 15 hr during normal pupal development. Electron microscopy shows that these fibers are present from 40 to 53 hr in bristles of wild-type flies but are absent in the null f36a mutant. The forked protein(s) thus appear to be an essential part of the bristle fibers. The phenotype of the f36a mutation can be rescued by a 13-kb fragment of the forked locus containing the coding regions for the 2.5, 1.9, and 1.1-kb transcripts, suggesting that the proteins encoded by the three large forked RNAs are dispensable during bristle development. Increasing the copy number of a P[w+,f+] construct containing the 13-kb fragment induces a hypermorphic bristle phenotype whose severity correlates with the number of copies of P[w+,f+] present. These results indicate that alterations in the ratios among the forked proteins, or between forked products and other components of the fiber, result in abnormal assembly of the fibrillar cytoplasmic structures necessary for bristle morphogenesis.

P element-induced gene conversion has been previously used to modify the white gene of Drosophila melanogaster in a directed fashion. The applicability of this approach of gene targeting in Drosophila melanogaster, however, has not been analyzed quantitatively for other genes. We took advantage of the P element-induced forked allele, f hd , which was used as a target, and we constructed a vector containing a modified forked fragment for converting f hd . Conversion frequencies were analyzed for this locus as well as for an alternative white allele, w eh812 . Combination of both P element-induced mutant genes allowed the simultaneous analysis of conversion frequencies under identical genetic, developmental, and environmental conditions. This paper demonstrates that gene conversion through P element-induced gap repair can be applied with similar success rates at the forked locus and in the white gene. The average conversion frequency at forked was 0.29%, and that at white was 0.17%. These frequencies indicate that in vivo gene targeting in Drosophila melanogaster should be applicable for other genes in this species at manageable rates. We also confirmed the homolog dependence of reversions at the forked locus, indicating that P elements transpose via a cut-and-paste mechanism. In a different experiment, we attempted conversion with a modified forked allele containing the su(Hw) binding site. Despite an increased sample size, there were no conversion events with this template. One interpretation (under investigation) is that the binding of the su(Hw) product prevents double-strand break repair.Gene targeting in eukaryotes through homologous recombination between exogenous DNA and its endogenous chromosomal homolog was first investigated with yeast cells (15,31) and subsequently successfully tested with mammalian cells (1,3,21). In Drosophila melanogaster, P element-mediated germ line transformation (34, 37) was used to introduce modified DNA fragments to integrate at quasirandom sites into the genome. However, the altered chromosomal position of the modified gene often leads to unwanted position effects. Furthermore, many genes are too large to be incorporated in a fully functional P element and must be studied in situ.More recently, directed modification of the white gene of Drosophila melanogaster has been carried out successfully (10). The technique depends on the presence of a P transposable element within or near the targeted gene. The mechanism involves a transposon-induced DNA double-strand break which subsequently is repaired by molecular processes of homologous recombination (for a review of P element-induced gap repair, see reference 22). The site of the double-strand break is precisely defined, and a template for repair can be designed in vitro. These studies suggested a new model for double-strand break repair, the synthesis-dependent strand annealing (SDSA) model (29), which may be of general relevance for homologous recombination in higher eukaryotes. According to the SDSA model, one or both ends of t...