Research Report 1992 of the Fischbach lab

Neurogenetic Analysis of Visual System Development in Drosophila melanogaster

Being our first report published in the www, it summarizes some of our key results about visual system development obtained in the past.

1. Compound eye development is independent from the optic lobe

The Heisenberg-Böhl procedure for simultaneous handling and sectioning of many flies permitted large mutant screens using brain morphology as the only selection criterion. One mutant identified by such a histological screen is disconnected (disco). In this mutant the larval optic nerve (Bolwig's nerve) often fails to innervate the brain during embryogenesis and the compound eye is often not connected to the optic lobe. Genetic mosaic analysis has shown that the genotype of the compound eye is not relevant for the development of this phenotype, i.e. in flies mosaic for disco, genetically wild type compound eyes can be unconnected to the optic lobe. Actually, the disco mutation has no apparent influence on compound eye development. The eye imaginal disc develops quite normally even in the absence of contacts of the retinula cells to the brain.

Fischbach K.-F. und Heisenberg M. (1984). Neurogenetics and Behaviour in Insects. J. exp. Biol. 112 , 65-93.

Steller H., Fischbach K.-F., und Rubin G.M. (1987). Disconnected : A locus required for neuronal pathway formation in the visual system of Drosophila. Cell 50, 1139-1153

2. Survival of retinula cells depends on connections to the optic lobe

Although optic lobe contacts are not required for compound eye development, they are necessary for maintenance of photoreceptors. In unconnected compound eyes of disco mutants photoreceptors start to degenerate a few days after eclosion. Genetic mosaic analysis has shown that this degeneration is not due to an autonomous requirement of gene function in the eyes. It is instead strictly dependent on their contacts to the optic lobe.

Campos A.R., Fischbach K.-F., Steller H. (1992). Survival of the photoreceptor neurons in the compound eye of Drosophila depends on connections with the optic ganglia. Development 114, 355-366

3. Optic lobe development depends on retinal fibre ingrowth

In completely eyeless flies no lamina can be detected anymore, and medulla and lobula complex are severely reduced by cell degeneration. Using genetic mosaic analysis it was shown that the effect of the compound eye on the optic lobe is indeed inductive. A reduction of retinal innervation leads to a cascade of degenerative events in genetically wild type optic lobes.

Fischbach K.-F. und Technau, G. (1984). Cell degeneration in the developing optic lobes of the sine oculis and small optic lobes mutants of Drosophila melanogaster. Dev. Biol. 104, 219-239

4. Topological maps develop in the absence of retinal innervation

One of the most intriguing problems of developmental biology is posed by topological maps of sensory fields in the brain. In the case of invertebrates a strong case has been made that the spatio- temporally ordered, sequential ingrowth of retinal fibres organizes the optic lobe into neuroommatidia. It has been proposed that neurons of the medulla cortex, which like lamina monopolar cells are produced by the outer optic anlage, may in a similar manner require contact with long visual receptor or lamina fibers for differentiation. In this scheme, the topological organization of the compound eye would be imposed on successive synaptic layers of the optic lobe by a cascade of inductive events initiated by the sequential ingrowth of retinal fibres. A detailed analysis of Golgi impregnated optic lobe rudiments of eyeless so-mutants has, however, shown that isotopic connections between the medulla, lobula and lobula plate rudiments form in the absence of retinal fibres. This result indicates that birth and axon formation of at least some of the neuronal types in the medulla and lobula complex are not triggered by retinal ingrowth. Furthermore, the es ablishment of a topological relationship among these neuropiles seems to be an optic lobe autonomous feature.

Fischbach K.-F. (1983). Neural cell types surviving congenital sensory deprivation in the optic lobe of Drosophila melanogaster. Dev. Biol. 95, 1-18

5. Visual axons "know" where to go

The phenotype of mutants with irregular axonal projections in the optic chiasms is telling with regard to mechanisms of axonal navigation in the optic lobe. Mutants of the irregular chiasm C -roughest (irreC-rst) locus are phenotypically and genetically best characterized. Two alleles affecting axonal projections have been isolated, one resulting from hybrid dysgenesis and the other from X-ray mutagenesis. Both were isolated due to a severe disorder of the second optic chiasm. Axon fascicles from the medulla to the lobula plate connect to their target region by projecting directly through the lobula neuropile instead of taking the second optic chiasm detour. In some severe cases the neuropiles of medulla, lobula and lobula plate seem to be partially fused. Most spectacular in our eyes, however, is a much more subtle, albeit related disorder of the first optic chiasm. One or occassionally several axon fascicles from the most posterior part of the lamina project to their normal target area in the anterior medulla via ectopic routes. The ectopic routes are variable, but in all cases the fascicles terminate in their appropriate target regions in the anterior medulla. Although the cellular mechanisms by which this phenotype develops are not yet fully elucidated, it clearly shows that neighbouring axon fascicles of the first optic chiasm are able to navigate independently from each other. irreC-rst mutants can also be used to illustrate that lamina axons recognize their target layer in the medulla irrespective of their direction of ingrowth. L3 fibres entering layer M3 ectopically from the proximal medulla form normal terminals.

Taken together our findings seem to exclude that retinotopic projections in the optic lobe of Drosophila are merely a result of spatiotemporal coordination of axon outgrowth. Younger axon fascicles do not have to be guided by older ones in order to reach their targets and the direction of ingrowth into the target region does not seem to be essential. The independent navigation of retinal axon fascicles stresses the importance of positional cues in the micro-environment, e.g. guidepost cells or molecular gradients.

Boschert U., Ramos R.G.P., Tix S., Technau G.M., and Fischbach K.-F. (1990). Genetic and Developmental Analysis of irreC, a genetic function required for optic chiasm formation in Drosophila. J. Neurogenetics 6, 153-171

6. Towards the molecular and cellular mechanisms underlying retinotopic map formation in Drosophila

The irreC-rst gene product is a novel member of the immunoglobulin superfamily that aligns over the entire length of the extracellular domain with the chicken axonal surface protein BEN. Both transmembrane proteins have 5 Ig- and no fibronectin domains and are dynamically expressed during development in subpopulations of neural and non-neural tissues. In the cerebellum, for example, BEN is specifically expressed on the climbing fibres during axonal growth and synaptogenesis. For the irreC-rst protein antibodies that reveal its expression pattern have recently been produced (in preparation). The irreC-rst gene product clearly affects axonal projections in the visual system of Drosophila, as is demonstrated by the mutant phenotypes. The fact that inspite of the erraneous routes the axons terminate in their approximate target region of the medulla indicates that axons use different and most likely combinations of cues en route. The identification of these cues requires a detailed characterization of cellular and extracellular components that interact with growing axons. In addition to the classical mutant approach, the enhancer trap technique can be used. This method is well suited for the isolation of cell type specific markers and provides easy access to the cloning of cell type specific genes, e.g. enhancer trap lines have been isolated that show reporter gene expression in a homologous set of large glial cells in both chiasms. In each chiasm these glial cells are aligned in a vertical row separating the axon fascicles of different horizontal planes. Their possible role in axon guidance is suggested by the disruption of their alignment in irreC mutants. The enhancer trap technology has also revealed position-specific expression patterns as well as gradients of gene expression in retina and optic lobe. On the background of the above mentioned information about the navigation of visual fibres, it will be intriguing to learn what role such position specific information plays during retinotopic map formation in the visual system of Drosophila.

Eule E., Tix S., Fröhlich A. and Fischbach K.-F. (1992). Glial cells of the optic lobe of Drosophila melanogaster. Verh. dt. Zool. Ges. 85.1, 63

Schneider T. and Fischbach K.-F. (1992). Position-specific lacZ-reporter gene expression in early embryogenesis and in the developing optic lobes of Drosophila. In Rhythmogenesis in neurons and networks. A 641. Eds. N. Elsner and D.W. Richter.

Ramos R.G.P., Igloi G.L. and Fischbach K.-F. (1992). A putative gene product of the irreC locus has characteristics of extracellular proteins involved in cell adhesion and cellular recognition. In Rhythmogenesis in neurons and networks. A 676. Eds. N. Elsner and D.W. Richter.

7. A molecular link between programmed cell death and axonal projection defects

(in preparation)

The irreC-rst gene is strongly expressed in the eye imaginal disc. The retinae of irreC UB883 and irreC 1R34, those alleles that were isolated due to axonal projection defects in the optic lobe, show subtle and variable defects in the wildtype retinae, irreC 1R34 being more severe than irreC UB883. The differences concern interommatidial cells (secondary and tertiary pigment cells and the cells of the bristle cell complex). The bristle complex e.g. has partially lost its contacts to primary pigment cells. The overall appearance of roughness of the eye though is far less severe than in rst CT which is the only known structural allele of the irreC-rst locus. It truncates the cytoplasmic domain of the protein. In rst CT supernumerary cells inbetween the ommatidia are not eliminated by the normally occurring programmed cell death (Wolff and Ready, Development 113, 825-839; 1991).

Our results suggest that during compound eye development irreC-rst is required for the right arrangement of interommatidial cells. Interestingly, integrity of the cytoplasmic domain is not required for normal optic lobe development.

Initially translated into html by Dr. C. Rees