The long term goal of these studies is to understand the presynaptic mechanisms that lead to transmitter release at fast chemical synapses. Our specific objective is to reconstruct in three dimensions the ultrastructure of the hair cell afferent synapse in order to test several aspects of our working model of this synapse's function. The hair cell afferent synapse is a ribbon-class synapse, characterized by an electron dense organelle of unknown function, the synaptic body, which appears to hover in the cytoplasm a vesicle's breadth above the active zone. Clear-core vesicles surround the synaptic body, but are also found in the cytoplasm, and associated with the plasma membrane at the active zone. What is the function of the synaptic body, and what is the functional significance of these various pools of vesicles? We will address these issues by determining the precise geometry of the active zone and synaptic body, and map the number and location of synaptic vesicles. By testing the activity-dependence of these measures, we will be able to integrate ultrastructure with ongoing physiological and mathematical modeling studies to build a more detailed understanding of the steps that lead to transmitter exocytosis. We have used electron tomography to reconstruct eleven afferent synapses, and two non-synaptic regions, in frog saccular hair cells. A thick section containing the region of interest was imaged in the IVEM at a series of tilt angles, and the micrographs digitized and back-projected to generate three-dimensional data sets from the two-dimensional projections. By tracing organelles in each plane in which they appeared, we have mapped the locations of nine synaptic bodies, 2573 vesicles, the plasma membranes, presynaptic density, and in some cases, mitochondria and endoplasmic reticulum. Traced organelles were then rendered to reveal their three-dimensional structures and relationships. A shell of vesicles surrounded the synaptic body, and vesicles were distributed uniformly and randomly across the surface, including the space between it and the plasma membrane. At several synapses, the plasma membrane followed the curve of the lower portion of the synaptic body, forming a space one to two vesicles wide between them, and creating a bulge in the cell's surface. In previous, conventional transmission electron micrographs of frog saccular hair cells, synaptic bodies always showed a round profile, arguing that they were spherical. To test this hypothesis, we asked whether spheres fit the synaptic bodies we reconstructed. Our data show that the synaptic body was well fit by a sphere when viewed transected along two perpendicular planes. Since all synaptic bodies were incomplete, it is possible that their unreconstructed portion were not spherical. This is unlikely, however, since each of nine reconstructed synaptic bodies was well fit by spheres, including two which were more than two thirds whole. From these fits, we estimate the average synaptic body diameter to be 470 nm (n = 9). In the cells we study, our capacitance measurements estimated that frog saccular hair cells could maintain exocytosis equivalent to the fusion of 500 vesicles per synapse per second for at least 2 seconds. We therefore counted vesicles at reconstructed synapses to determine the possible ultrastructural basis for this rate. We counted approximately 100 to 200 synaptic body-associated vesicles, which, after correcting for the missing portion of the sphere, gave an average of 376 (n = 9) vesicles per whole synaptic body. Vesicles were packed at 55% of the synaptic body's carrying capacity. Synaptic body-associated vesicles could therefore only account for 752 ms of exocytosis, or 1.37 seconds if all synaptic bodies were loaded at carrying capacity. We also counted outlying vesicles in the reconstructed synapses, and after subtracting the volumes filled by the synaptic body, its associated vesicles, mitochondria, large membranous compartments, and the postsynaptic cytoplasm, we calculated that outlying vesicles occupied 4.2% of available cytoplasm (0.97 um3) in the vicinity of the synapse. To determine whether vesicles were formed at, or actively delivered to synapses, we compared this concentration to the background vesicle density in non-synaptic areas. We used tomography to reconstruct a non-synaptic region of hair cell cytoplasm adjacent to the plasma membrane at an unknown distance from any synapse, but in the basal portion of the cell where synapses are most common. Here, 91 vesicles occupied 0.3% of the available cytoplasm (0.96 um3). Since outlying vesicle concentration was higher in the neighborhood of the synapse, vesicles are either manufactured locally, or translocated toward synapses. Using the background vesicle concentration (0.9% of available volume), and the estimated volumes of three hair cells, we calculated that a hair cell contains approximately 600 000 clear-core vesicles throughout its cytoplasm. If there are 20 afferent synapses per hair cell, then there is a pool of about 30 000 vesicles per synapse, which is two to three times as large as in bipolar terminals. Since some portion of these vesicles will serve non-synaptic functions, and because capacitance measurements estimate that continuous exocytosis may endure for 10 times longer in hair cells than in bipolar, then the larger vesicle pool in hair cells cannot wholly account for the different staminas of exocytosis in these two cells. Miniature postsynaptic potentials (minis), the basis for the quantal hypothesis of transmitter secretion, vary in amplitude at all synapses where they have been observed, and variance in vesicle sizes could account for the mini distribution. Although vesicle diameter distributions from transmission electron micrographs have been reported for many other synapses, including ribbon synapses, data are often pooled between synapses, or are not corrected for sampling biases. Consequently, we have exploited electron tomography to measure many vesicles at single synapses, at high resolution. Dissecting reconstructed volumes in planes one voxel thick allowed us to serial section the synapse with an order of magnitude better resolution than would have been possible with physical sections. We fit the polygon traced at each vesicle's equator to a circle of equal area, and plotted the distribution of equivalent diameters for vesicles at a single synapse, and at all eight synapses. At a single synapse, the synaptic body-associated vesicles (38.8 +/-3.4 nm, n = 139) and outlying vesicles (38.3 +/-3.5 nm, n = 208) were not significantly different, which was also true at
Showing the most recent 10 out of 384 publications
