Tectorotundal connections in turtles: An electron microscopic tracing and GABA-immunocytochemical study

Abstract

The nucleus rotundus of the turtles Emys orbicularis and Testudo horsfieldi was analysed by axonal tracing methods and post-embedding GABA immunocytochemistry. After injections of horseradish peroxidase or biotinylated dextran amine into the optic tectum, electron microscopic observations showed that the vast majority of ipsilateral tectorotundal axon terminals were small in size, had smooth contours and contained small, round, densely packed synaptic vesicles. These terminals were GABA-immunonegative, often gathered in clusters, and established asymmetrical synaptic contacts with either small- or medium-sized GABA-negative dendritic profiles and with GABA-immunoreactive (GABA-ir) dendrites, which did not contain synaptic vesicles. Occasional GABA-ir-labelled axon terminals were observed; these may arise from the rare GABAergic neurons in the central tectal layer, or from neurons in the ventral pretectal nucleus, which projects both to the optic tectum and nucleus rotundus. In addition to tracer-labelled axon terminals, we observed both GABA-negative and GABA-ir cell bodies and dendrites also labelled by the tracer. No GABA-ir presynaptic dendritic profiles containing synaptic vesicles were observed. The existence in reptiles of reciprocal connections between the nucleus rotundus and the optic tectum as a phylogenetically ancient feedback system is discussed.