We report a comparative immunofluorescence and immunoblotting study of GFA protein, the subunit of glial filaments, in nonmammalian vertebrates. The study was conducted with polyclonal antibodies raised to human and shark antigen and with monoclonal antibodies isolated from mice immunized with chicken and bovine antigen. With the exception of cyclostomes, glial filaments appeared remarkably conserved in vertebrate phylogeny, both with respect to the molecular weight and immunoreactivity of their protein subunit. In most species, the antibodies decorated a single band in brain, spinal cord, and optic nerve extracts by the immunoblotting procedure. This band had the same molecular weight in the different CNS regions. With the exception of the turtle, species differences in the molecular weight of the band were not greater than those observed among mammalian vertebrates (human, bovine, and rat). However, there were some exceptional findings in fish. In goldfish and trout brain and spinal cord extracts, the antibodies decorated with the same intensity two bands. In accordance with previous immunofluorescence findings, goldfish optic nerve extracts were negative by the immunoblotting procedure. In four fishes (sea bass, tautog, trout, and scup), optic nerves reacted with the antibodies. However, the band decorated by the antibodies was higher in molecular weight than that obtained from brain and spinal cord extracts. Glial fibers were demonstrated by immunofluorescence in the brain, spinal cord, optic nerve, and retina of most species studied. In amphibia immunofluorescent structures were comparatively few, probably accounting for the negative results by immunoblotting. A comparative immunohistological study of the cerebellum showed the presence of perpendicular glial fibers in the molecular layer of most species examined. Birds and amphibia were different in this respect. Bergmann glia in chicken were GFA negative. In the frog and the toad, immunofluorescent fibers in the molecular layer of the cerebellum were haphazardly oriented. Ependymal radial glia was GFA-negative in the cerebellum of subavian vertebrates. Antisera raised in rabbit to shark GFA protein reacted with the same bovine GFA fragments recognized by polyclonal and monoclonal antibodies raised to human and bovine antigens, respectively, i.e., 30-kDa N-bromosuccinimide fragment (tryptophan cleavage); 35-kDa 2-nitro-5-thiocyanobenzoic acid fragment (cysteine cleavage); 18-kDa cyanogen bromide fragment (methionine cleavage). Conversely, the chicken GFA monoclonal antibodies selected for this study only reacted with noncleaved protein.

Gap junctional proteins are important components of signaling pathways required for the development and ongoing functions of all animal tissues, particularly the nervous system, where they function in the intracellular and extracellular exchange of small signaling factors and ions. In animals whose genomes have been sufficiently sequenced, large families of these proteins, connexins, pannexins, and innexins, have been found, with 25 innexins in the nematode Caenorhabditis elegans Starich et al. (Cell Commun Adhes 8: 311-314, 2001) and at least 37 connexins in the zebrafish Danio rerio Cruciani and Mikalsen (Biol Chem 388:253-264, 2009). Having recently sequenced the medicinal leech Hirudo verbana genome, we now report the presence of 21 innexin genes in this species, nine more than we had previously reported from the analysis of an EST-derived transcriptomic database Dykes and Macagno (Dev Genes Evol 216: 185-97, 2006); Macagno et al. (BMC Genomics 25:407, 2010). Gene structure analyses show that, depending on the leech innexin gene, they can contain from 0 to 6 introns, with closely related paralogs showing the same number of introns. Phylogenetic trees comparing Hirudo to another distantly related leech species, Helobdella robusta, shows a high degree of orthology, whereas comparison to other annelids shows a relatively low level. Comparisons with other Lophotrochozoans, Ecdyzozoans and with vertebrate pannexins suggest a low number (one to two) of ancestral innexin/pannexins at the protostome/deuterostome split. Whole-mount in situ hybridization for individual genes in early embryos shows that ∼50% of the expressed innexins are detectable in multiple tissues. Expression analyses using quantitative PCR show that ∼70% of the Hirudo innexins are expressed in the nervous system, with most of these detected in early development. Finally, quantitative PCR analysis of several identified adult neurons detects the presence of different combinations of innexin genes, a property that may underlie the participation of these neurons in different adult coupling circuits.