Source: Dissertation Abstracts International, Volume: 60-06, Section: B, page: 2490.

Director: Robert J. Wyman.

The Passover gene codes for structural molecules of gap junctions in Drosophila. This thesis describes the complex set of alternatively spliced transcripts from Passover. The functions of some of the family members were tested in cultured cell lines and in Drosophila, and some of the structural and biochemical properties of the Passover proteins were investigated.

I begin by describing the molecular cloning of new Passover transcripts and their expression patterns. The transcripts share 3 ' exons, but each has its own 5' exon. The transcripts are arrayed as a series in the genomic DNA stretching over 60 Kb. The same coding region is included in three of the transcripts, each of which has its own unique 5' non-coding sequence. Three different proteins are encoded, PasV, PasN and PasN+16.

The different transcripts are expressed in a wide variety of locations in the nervous system and in non-neural tissues. The expression pattern for the different transcripts sometimes overlap and thus some tissues express more than one transcript, suggesting the possibility of heteromeric hemi-channels. Within the adult CNS, these transcripts have an expression pattern that is restricted to the Giant Fiber System (GFS). The pasN1 transcript, previously defined as active in the Giant Fiber, is not, in fact, expressed in that cell. Instead, I find that another transcript, pasN3, and maybe also pasN4, is expressed in the GFS. Two other transcripts, pasN1 and pasN2 are expressed in the optic lamina, but not elsewhere in the CNS. This expression pattern explains the neurophysiological and behavioral defects in escape exhibited in mutants of pas.

The functional role of the Passover family members was investigated by expressing them in Drosophila using the UAS-GAL4 system, and by expressing them in cultured cell lines. In Drosophila , I found that all the tested family members were functional.

The different rescue capabilities of the different transcripts indicate that the transcripts are not equivalent. It appears that pasV is able to form homomeric and homotypic channels, whereas PasN and PasN+16 seem to require heteromeric or heterotypic partners.

Using the PasV protein as a model, some of the basic properties of invertebrate gap junction proteins were investigated. The PasV protein is an integral membrane protein with an intracellularly localized C-terminus. It appears to have two isoforms, which may arise from differential post-translational modification. These two isoforms have different detergent solubility.

Even though their physiological properties are quite similar, gap junctions seem to be made from completely different proteins in vertebrates and invertebrates. Invertebrates do not appear to make connexins and vertebrates do not appear to have pas-homologous proteins. Passover, however, can form junctions in vertebrates and I demonstrate that vertebrate connexins probably form junctions in Drosophila. (Abstract shortened by UMI.)