Transmission at a central inhibitory synapse. III. Ultrastructure of physiologically identified and stained terminals.

Abstract

Studies were conducted on neurons inhibitory to the teleost Mauthner cell (M-cell) at both light microscopic and EM levels. Neurons belonging to the goldfish and tench M-cell collateral inhibitory network were identified physiologically and intracellularly stained by iontophoretically injected horseradish peroxidase (HRP). The cells appeared bipolar. Their axons issued 2 groups of terminals presynaptic to the M-cell; one of them penetrated within the M-cell''s axon cap and terminated on this neuron as unmyelinated club endings (UCE), while the 2nd group remained outside this region and issued small vesicle boutons (SVB). Transverse sections of HRP-filled terminals revealed that both types exhibit typical aspects of Gray type II synapses with distinct pre- and postsynaptic differentiations and contain pleiomorphic vesicles. A morphometric analysis of adjacent unlabeled UCE and SVB arising from both collateral and commissural cells indicated that the length of their respective apposition zones, the size of their postsynaptic differentiations and the mean or maximum number of PDP [presynaptic dense projections] encountered at the level of their presynaptic differentiations were all identical. The 2 populations of synaptic boutons cannot be distinguished on the basis of such criteria. This finding strengthens the concept, developed on physiological grounds, that UCE and SVB are functionally equivalent. In transverse sections stained with ethanolic phosphotungstic acid (E-PTA) the extracellular space outlining individual terminal boutons was darkly stained. Individual terminals visualized showed that 95% of the endings contained only 1 presynaptic grid. A correspondence between inhibitory post synaptic potentials (IPSP) and the number of presynaptic grids present in the terminals of the releasing interneurons was observed. E-PTA-stained transverse sections also showed that the PDP are organized in regular triagonal arrays linked at their base by thin filamentous cross bridges, a disposition typical of presynaptic grids in the vertebrate CNS. Calculations of PDP size and center-to-center distance yielded values that were close to those reported in HRP-filled terminals, which confirms that the enzyme allows a reliable analysis of synaptic ultrastructure. En face views of both unlabeled and E-PTA-strained terminals indicated that synaptic vesicles occupy the free space between PDP so that each of the latter is surrounded by 6 vesicles. The number of vesicles present in typical presynaptic complexes and presumably available for transmitter release was .apprx. 44/ellipsoid grid and 62/annulated grid. A comparison of the geometrical features and the physiological properties characterizing the dense bands of neuromuscular junctions and the presynaptic grids of central synapses was consistent with the concept that these elements constitute morphological synaptic unit functioning under the same constraints. These observations provide a morphological substrate for the concept, derived from binomial analysis of fluctuating IPSP that following a presynaptic impulse, the contents of only 1 synaptic vesicle is released by each terminal bouton. A simple model is proposed to account for the mechanism by which exocytosis of 1 vesicle inhibits the discharge of its neighbors throughout the whole presynaptic grid.