In vivo functional domain analysis of the IrreC-rst adhesion molecule: Differential role of individual immunoglobulin domains in eye and optic lobe development

CHRISTIAN REITER* , JOHANNES KORN , AND KARL-FRIEDRICH FISCHBACH

Institut für Biologie III, Albert-Ludwigs-Universität Freiburg, Schänzlestr. 1, D-79104 Germany

INTRODUCTION: Irregular chiasmC-roughest (IrreC-rst) is an adhesion molecule of the immunoglobulin superfamily that was originally identified due to its involvement in processes of axonal pathfinding (1). Previous investigation of its function has shown that it acts as a homophilic adhesion molecule in cell culture assays (2), directs axonal projections in the optic chiasms (3), and is necessary for initiation of apoptosis during pupal development of the eye imaginal disc (4). Analysis of the wild-type expression pattern, mutant phenotypes, and misexpression studies using the Gal4/UAS system indicate heterophilic interactions with an as yet unidentified ligand in the eye imaginal disc (5), while homophilic IrreC-rst interactions appear sufficient for at least part of the protein´s function in axonal guidance. We have begun to investigate wether the multiple developmental functions of IrreC-rst can be assigned to individual immunoglobulin domains of the protein.

METHODS: We have misexpressed partially deleted irreC-rst constructs using the Gal4-UAS system. Phenotypical effects of misexpressing deletion constructs or wild-type protein were compared. Misexpression in cone cells was utilized as a bioassay for the ability of truncated constructs to affect cell death in the eye. Enhancer trap lines targeting the optic chiasms were used to assay remaining effects on axonal guidance. A monoclonal antibody directed against the N-terminus of the molecule was used to specifically mark only wild-type protein, while truncated versions were detected with a polyclonal serum (A).

RESULTS AND DISCUSSION: Targeting of both full-length and DIg1 to cone cells results in a rough eye phenotype. Investigation on a cellular level, however, shows drastic differences: the main effect of full-length protein is on the positioning of interommatidial cells (IOC). DIg1 does not affect IOC, but instead induces severe defects of cone cell formation (B). On the protein level, full-length protein in cone cells affects localization of wild-type protein in IOC, while DIg1 does not (C).

Thus, DIg1 seems to be incapable of competing with wt protein for the binding activity present on primary pigment cells, and the developmental function of the truncated protein is drastically altered.

In the optic lobe, DIg1 induces misexpression phenotypes that are similar to, but generally weaker than, those caused by full-length protein, e.g. assembly of ectopic neural plexi (D). Therefore it can be concluded that the most N-terminal Ig domain is essential for the interactions driving cell sorting in the pupal eye imaginal disc and initiating IOC apoptosis, but dispensable for the deliverance of axonal guidance cues. Furthermore, developmental functions potentially mediated by the first four Ig domains seem to be masked by the presence of the first domain. Its removal enables IrreC-rst to exert influences on cell fate decisions that are not affected in mutants or transformants misexpressing full-length protein. We will present further data concerning deletions of other domains of IrreC-rst.

e-mail contact (Christian Reiter)