Cell lineage analysis of adult fly brains using the flippase / FRT and GAL4 / UAS systems

KEI ITO (P), WAKAE AWANO AND DAISUKE YAMAMOTO

ERATO, Yamamoto Behaviour Genes Project, at Mitsubishi Kasei Institute of Life Sciences, 194 Machida-shi, Minami-Ooya 11, Tokyo, JAPAN
tel: (81) 427-21-2334, fax: (81) 427-21-2850, email: itokei@fly.erato.jrdc.go.jp

The central nervous system consists of numerous neural circuits, which form various substructures in the neuropile. In Drosophila, the neurons in the central brain are produced by an average of 85 neuroblasts per hemisphere. At least one of the neuropile substructures, the mushroom body, is known to be made by the four distinct neuroblasts. In other cases, however, the relationship between the cell lineage and the functional substructure is virtually unknown. Do the progeny of a single neuroblast all form a similar circuit? Or do they contribute to diverse neuropile structures?

Cell lineage analysis of the embryonic nervous system has been performed successfully by transplanting single neuroblasts or applying tracers (HRP, DiI, etc.) to single cells. These methods, however, are not easily applicable to study adult brains. In this study, we addressed this question by using the combination of flippase / FRT and GAL4 / UAS systems with the UAS-tau axonal marker. We made flies that carry hsp-FLP (Golic and Lindquist, Cell 59: 499-509), AyGAL4 (Hiromi, unpublished) and UAS-tau (Ito, Schneuwly et al., in preparation). The AyGAL4 contains an "FRT - hsp termination signal - yellow - FRT" cassette between the actin promoter and the GAL4 gene. Upon mild heat shocking, the flippase removes the FRT cassette in a small subset of neuroblasts, starting GAL4 expression in these cells and their progeny. GAL4 then activates the expression of bovine tau; the resulting protein is actively transported to the distal ends of neuronal fibres, which can be revealed with the anti-Tau monoclonal antibody.

The number of active postembryonic neuroblasts reaches its maximum between the early third larval instar and the first-day pupal stages. To induce postembryonic clones, we therefore heat shocked larvae at early third instar. In the preparations fixed at the end of the larval stage, the neurons of each labelled clone form a tightly packed cluster. A single neuroblast is found labelled in each clone, confirming that the labelling is due to the flipping reaction which occurred in the stem cell. Each clone has only one fibre bundle projecting towards the neuropile, in which all the cell body fibres (neurites) fasciculate.

After metamorphosis, the cluster of labelled cells in a clone becomes less tightly packed. In most cases, however, it still maintains a single bundle of cell body fibres, which runs into the neuropile until the fibres reach their first destination. We found many clones in which all the cells innervate single neuropile substructures, e.g. antennal lobe, mushroom body, lateral horn, ventrolateral protocerebrum, and great commisure. Within the substructure, however, the innervation pattern of all the cells is not the same; the cells of a mushroom body clone innervate all the three lobes (alpha, beta and gamma), for example. We also found clones that contribute to two or three substructures stereotypically, but virtually no clone showed random innervation of multiple substructures. Thus, it is likely that the cell lineage plays an important role in defining the cell populations that contribute to a particular neuropile substructure.