Search

The first synapse of the visual pathway is formed by photoreceptors, horizontal cells and bipolar cells. While ON bipolar cells invaginate into the photoreceptor terminal and form synaptic triads together with invaginating horizontal cell processes, OFF bipolar cells make flat contacts at the base of the terminal. When horizontal cells are ablated during retina development, no invaginating synapses are formed in rod photoreceptors. However, how cone photoreceptors and their synaptic connections with bipolar cells react to this insult, is unclear so far. To answer this question, we specifically ablated horizontal cells from the developing mouse retina. Following ablation around postnatal day 4 (P4)/P5, cones initially exhibited a normal morphology and formed flat contacts with OFF bipolar cells, but only few invaginating contacts with ON bipolar cells. From P15 on, synaptic remodeling became obvious with clustering of cone terminals and mislocalized cone somata in the OPL. Adult cones (P56) finally displayed highly branched axons with numerous terminals which contained ribbons and vesicular glutamate transporters. Furthermore, type 3a, 3b, and 4 OFF bipolar cell dendrites sprouted into the outer nuclear layer and even expressed glutamate receptors at the base of newly formed cone terminals. These results indicate that cones may be able to form new synapses with OFF bipolar cells in adult mice. In contrast, cone terminals lost their invaginating contacts with ON bipolar cells, highlighting the importance of horizontal cells for synapse maintenance. Taken together, our data demonstrate that early postnatal horizontal cell ablation leads to differential remodeling in the cone pathway: whereas synapses between cones and ON bipolar cells were lost, new putative synapses were established between cones and OFF bipolar cells. These results suggest that synapse formation and maintenance are regulated very differently between flat and invaginating contacts at cone terminals.

The magnetic compass of birds seems to be based on light-dependent radical-pair processes in the eyes. Cryptochromes are currently the only candidate proteins known in vertebrates that may serve as the primary radical-pair-based magnetoreceptor molecules. Previous immunohistochemical studies have suggested that cryptochrome 1a (Cry1a) is localised in the photoreceptor outer segments of the ultraviolet/violet (UV/V) cones, and it has been claimed that differences in Cry1a antibody staining intensities show that Cry1a is activated by light and that this should make Cry1a the most likely magnetoreceptive candidate molecule. Here, we present an independent study of Cry1a distribution within retinae of several bird species, ranging from non-migratory domestic chicken and rock pigeon to night-migratory passerines, using both the previously used antibody and two newly generated antibodies, one against the same epitope as the originally used antibody and one against a different epitope of Cry1a. We confirm the UV/V cone outer segment localisation of Cry1a in all the tested bird species. In some stainings, we found Cry1a immunoreactivity as a distinct punctate pattern throughout the whole length of the UV/V cone outer segments. These dots with a diameter of around 170 nm might suggest that many Cry1a molecules accumulate in distinct spots in the UV/V cone outer segments. However, we did not see any notable difference in Cry1a immunoreactivity between light- and dark-adapted retinae. We find no evidence whatsoever that a C-terminal antibody against Cry1a labels only a light-activated form of the Cry1a protein.

In the central nervous system, neuronal processing relies on the precisely orchestrated formation of synapses during development. The first synapse of the visual system is a triad synapse, comprising photoreceptors, horizontal cells and bipolar cells. During the second postnatal week, the axon terminal processes of horizontal cells invaginate rod spherules, followed by rod bipolar cell dendrites. Both elements finally oppose the synaptic ribbon (the release site of glutamate). However, it has not been fully elucidated whether horizontal cells are essential for rod bipolar cell dendrites to find their way into the rod terminal. In the present study, we investigated this question by specifically ablating horizontal cells from the early postnatal mouse retina. We monitored the formation of the rod-to-rod bipolar cell synapse during retinal maturation until postnatal day 21. Based on quantitative electron microscopy, we found that without horizontal cells, the dendrites of rod bipolar cells never entered rod terminals. Furthermore, rods displayed significantly fewer and shorter presynaptic ribbons, suggesting that glutamate release is decreased, which coincided with significantly reduced expression of postsynaptic proteins (mGluR6, GPR179) in rod bipolar cells. Collectively, our findings uncover that horizontal cells are indeed necessary guideposts for rod bipolar cells. Whether horizontal cells release diffusible guidance cues or provide structural guidance by expressing specific cell adhesion molecules remains to be seen.

Neuronal gap junctions formed by connexin36 (Cx36) and chemical synapses share striking similarities in terms of plasticity. Ca2+/calmodulin-dependent protein kinase II (CaMKII), an enzyme known to induce memory formation at chemical synapses, has recently been described to potentiate electrical coupling in the retina and several other brain areas via phosphorylation of Cx36. The contribution of individual CaMKII isoforms to this process, however, remains unknown. We recently identified CaMKII-β at electrical synapses in the mouse retina. Now, we set out to identify cell types containing Cx36 gap junctions that also express CaMKII-β. To ensure precise description, we first tested the specificity of two commercially available antibodies on CaMKII-β-deficient retinas. We found that a polyclonal antibody was highly specific for CaMKII-β. However, a monoclonal antibody (CB-β-1) recognized CaMKII-β but also cross-reacted with the C-terminal tail of Cx36, making localization analyses with this antibody inaccurate. Using the polyclonal antibody, we identified strong CaMKII-β expression in bipolar cell terminals that were secretagogin- and HCN1-positive and thus represent terminals of type 5 bipolar cells. In these terminals, a small fraction of CaMKII-β also colocalized with Cx36. A similar pattern was observed in putative type 6 bipolar cells although there, CaMKII expression seemed less pronounced. Next, we tested whether CaMKII-β influenced the Cx36 expression in bipolar cell terminals by quantifying the number and size of Cx36-immunoreactive puncta in CaMKII-β-deficient retinas. However, we found no significant differences between the genotypes, indicating that CaMKII-β is not necessary for the formation and maintenance of Cx36-containing gap junctions in the retina. In addition, in wild-type retinas, we observed frequent association of Cx36 and CaMKII-β with synaptic ribbons, i.e., chemical synapses, in bipolar cell terminals. This arrangement resembled the composition of mixed synapses found for example in Mauthner cells, in which electrical coupling is regulated by glutamatergic activity. Taken together, our data imply that CaMKII-β may fulfill several functions in bipolar cell terminals, regulating both Cx36-containing gap junctions and ribbon synapses and potentially also mediating cross-talk between these two types of bipolar cell outputs.