The retinal photoreceptor ribbon synapse is a chemical synapse structurally and functionally specialized for the tonic release of neurotransmitter. It is characterized by the presynaptic ribbon, an electron-dense organelle at the active zone covered by hundreds of synaptic vesicles. In conventional synapses, dense-core transport vesicles carrying a set of active zone proteins are implicated in early steps of synapse formation. In photoreceptor ribbon synapses, synaptic spheres are suggested to be involved in ribbon synapse assembly, but nothing is known about the molecular composition of these organelles. With light, electron, and stimulated emission depletion microscopy and immunocytochemistry, we investigated a series of presynaptic proteins during photoreceptor synaptogenesis. The cytomatrix proteins Bassoon, Piccolo, RIBEYE, and RIM1 appear early in synaptogenesis. They are transported in nonmembranous, electron-dense, spherical transport units, which we called precursor spheres, to the future presynaptic site. Other presynaptic proteins, i.e., Munc13, CAST1, RIM2, and an L-type Ca(2+) channel alpha1 subunit are not associated with the precursor spheres. They cluster directly at the active zone some time after the first set of cytomatrix proteins has arrived. By quantitative electron microscopy, we found an inverse correlation between the numbers of spheres and synaptic ribbons in the postnatally developing photoreceptor synaptic terminals. From these results, we suggest that the precursor spheres are the transport units for proteins of the photoreceptor ribbon compartment and are involved in the assembly of mature synaptic ribbons.

Presynaptic mutations leading to reduced photoreceptor-to-bipolar cell signaling are associated with disturbances in protein expression within postsynaptic dendrites. Moreover, detection of Cacna1s and Serca2 in ON-bipolar cell dendrites in wild-type animals suggests a putative role in regulation of postsynaptic Ca(2+) flux.

How size and shape of presynaptic active zones are regulated at the molecular level has remained elusive. Here we provide insight from studying rod photoreceptor ribbon-type active zones after disruption of CAST/ERC2, one of the cytomatrix of the active zone (CAZ) proteins. Rod photoreceptors were present in normal numbers, and the a-wave of the electroretinogram (ERG)-reflecting their physiological population response-was unchanged in CAST knock-out (CAST Ϫ/Ϫ ) mice. Using immunofluorescence and electron microscopy, we found that the size of the rod presynaptic active zones, their Ca 2ϩ channel complement, and the extension of the outer plexiform layer were diminished. Moreover, we observed sprouting of horizontal and bipolar cells toward the outer nuclear layer indicating impaired rod transmitter release. However, rod synapses of CAST Ϫ/Ϫ mice, unlike in mouse mutants for the CAZ protein Bassoon, displayed anchored ribbons, normal vesicle densities, clustered Ca 2ϩ channels, and essentially normal molecular organization. The reduction of the rod active zone size went along with diminished amplitudes of the b-wave in scotopic ERGs. Assuming, based on the otherwise intact synaptic structure, an unaltered function of the remaining release apparatus, we take our finding to suggest a scaling of release rate with the size of the active zone. Multielectrode-array recordings of retinal ganglion cells showed decreased contrast sensitivity. This was also observed by optometry, which, moreover, revealed reduced visual acuity. We conclude that CAST supports large active zone size and high rates of transmission at rod ribbon synapses, which are required for normal vision.

Immunocytochemical discrimination of distinct bipolar cell types in the mouse retina is a prerequisite for analyzing retinal circuitry in wild-type and transgenic mice. Here we demonstrate that among the more than 10 anatomically defined mouse bipolar cell types, type 4 bipolar cells are specifically recognized by anti-calsenilin antibodies. Axon terminals in the inner plexiform layer are not readily identifiable because calsenilin is also expressed in a subset of amacrine and ganglion cells. In contrast, in the outer plexiform layer calsenilin immunoreactivity allows the analysis of photoreceptor to type 4 bipolar cell contacts. A dense plexus of calsenilin-positive dendrites makes several basal contacts at cone pedicles. An individual calsenilin-positive bipolar cell contacts five to seven cones. In addition, some calsenilin-positive dendrites contact rod photoreceptors. On average we counted 10 rod spherule contacts per type 4 bipolar cell, and approximately 10% of rods contacted type 4 bipolar cells. We suggest that type 4 bipolar cells, together with the recently described type 3a and b cells, provide an alternative and direct route from rods to OFF cone bipolar cells. In the Bassoon DeltaEx4/5 mouse, a mouse mutant that shows extensive remodeling of the rod system including sprouting of horizontal and rod bipolar cells into the outer nuclear layer due to impaired synaptic transmission, we found that in addition mixed-input (type 3 and 4) OFF bipolar cells sprout to ectopic sites. In contrast, true cone-selective type 1 and 2 OFF cone bipolar cells did not show sprouting in the Bassoon mouse mutant.

Knowledge about the plastic and regenerative capacity of the retina is of key importance for therapeutic approaches to restore vision in patients who suffer from degenerative retinal diseases. In the retinae of mice, mutant for the presynaptic scaffolding protein Bassoon, signal transfer at photoreceptor ribbon synapses is disturbed due to impaired ribbon attachment to the active zone. In a long-term study we observed, with light and electron microscopic immunocytochemistry and electroretinographic recordings, two overlapping events in the Bassoon mutant retina, i.e. loss of photoreceptor synapses in the outer plexiform layer, and structural remodeling and formation of ectopic photoreceptor synapses in the outer nuclear layer, a region usually devoid of synapses. Formation of ectopic synaptic sites starts around the time when photoreceptor synaptogenesis is completed in wild-type mice and progresses throughout life. The result is a dense plexus of ectopic photoreceptor synapses with significantly altered but considerable synaptic transmission. Ectopic synapse formation is led by the sprouting of horizontal cells followed by the extension of rod bipolar cell neurites that fasciculate with and grow along the horizontal cell processes. Although only the rod photoreceptors and their postsynaptic partners show structural and functional remodeling, our study demonstrates the potential of the retina for long-lasting plastic changes.