Ribbon synapses are specialized chemical synapses found in the retina, inner ear, and pineal gland (Matthews and Fuchs, 2010; Moser et al., 2020). These synapses are continuously active and possess specialized active zones that allow fast, precisely timed signaling for prolonged periods of time. Exocytosis of synaptic vesicles occurs at the active zones close to the voltage-gated Ca²⁺-channels (Cav-channels) (Zenisek et al., 2000; Grabner and Moser, 2021). The active zones of ribbon synapses associate with large, electron-dense presynaptic specializations, the synaptic ribbons, that tether and deliver large numbers of synaptic vesicles to the active zone similar to an assembly line (Vaithianathan et al., 2019; Joselevitch and Zenisek, 2020). The ribbon-associated docked vesicles close to the active zone represent the vesicle pool that fuses with the fastest release kinetics (Moser et al., 2020; Grabner and Moser, 2021). Synaptic vesicles associated with more distant parts of the ribbon away from the active zone refill vacated release sites and promote continuous signaling (Zenisek et al., 2000; Vaithianathan et al., 2019; Joselevitch and Zenisek, 2020).

Synaptic ribbons are large presynaptic structures with a complex three-dimensional (3D) shape (Moser et al., 2020). The 3D shape and size of the ribbon are relevant for synaptic transmission/signaling because the ribbon geometry determines the pool size of ribbon-associated vesicles (Moser et al., 2020; Kesharwani et al., 2021). The photoreceptor synapse as the first synapse in the visual system transmits light stimuli to the inner retina. In the mouse retina, rod photoreceptors represent the major type of photoreceptors (95% of photoreceptors; Moser et al., 2020). Rod photoreceptor synapses are built in a fairly uniform manner. They contain a single, large active zone with a single large, horseshoe-shaped synaptic ribbon (Schmitz, 2009; Moser et al., 2020). Due to their large size and conspicuous horseshoe-shaped morphology, rod ribbons are particularly suitable for imaging analyses. In the inner retina, bipolar cell terminals possess multiple active zones with smaller synaptic ribbons that are morphologically and functionally more diverse than rod photoreceptor synapses in the outer retina (Moser et al., 2020).

Despite differences in size and shape, all ribbons contain the ribbon-specific protein RIBEYE (Schmitz et al., 2000; Moser et al., 2020), the core protein of all synaptic ribbons (Schmitz et al., 2000; Maxeiner et al., 2016; Becker et al., 2018; Jean et al., 2018). RIBEYE is evolutionarily conserved in vertebrates, and also serves as a major building block of synaptic ribbons in zebrafish (Lv et al., 2016; Sheets et al., 2017). Structurally, RIBEYE consists of a unique N-terminal, proline-rich A-domain, and a carboxyterminal B-domain that is identical (except for the N-terminal 20 amino acids) to the nuclear transcriptional co-repressor CtBP2 (Schmitz et al., 2000; Piatigorski, 2001). The RIBEYE B-domain/CtBP2 emerged from a family of D isomer-specific 2-hydroxyacid dehydrogenases (Goldberg et al., 1994). RIBEYE has multiple binding sites for other RIBEYE proteins both in the A- and B-domain of RIBEYE (Magupalli et al., 2008; Madison et al., 2013; Bellesis et al., 2018; Jecrois et al., 2021). However, the contribution of the RIBEYE A- and B-domains to the formation of synaptic ribbons remains unclear. Based on cell transfection experiments, the RIBEYE A-domain appears to serve mainly as a structural component of the synaptic ribbon scaffold (Schmitz et al., 2000; Magupalli et al., 2008), but it is unknown whether the RIBEYE A-domain alone is sufficient for the assembly of synaptic ribbons in-situ or whether this process also depends on RIBEYE B-domain.

In order to understand whether RIBEYE B-domain plays a role in the assembly of synaptic ribbons, we made use of RIBEYE^KI knockin (KI) mice (Maxeiner et al., 2016). In these KI mice, RIBEYE B-domain has been replaced by GCaMP3 in the KI allele, thus making these animals an ideal tool to study the role of the RIBEYE B-domain. We found that in the absence of the RIBEYE B-domain, synaptic ribbons are not assembled and the RIBEYE A-domain/GCaMP3 fusion protein is destabilized. Thus, the RIBEYE B-domain is essential for the assembly of synaptic ribbons.

In the current study, we analyzed RIBEYE knockin (KI) mice, RBE^KI (Maxeiner et al., 2016) to study whether RIBEYE B-domain has a role in the assembly of synaptic ribbons. In the RIBEYE KI mice, RIBEYE B-domain has been replaced by the cDNA of GCaMP3 in the knockin (KI) allele (Maxeiner et al., 2016), making these KI mice an ideal model to study the function of the RIBEYE B-domain.

First, we checked whether the overall organization of the RBE^KI/KI retina is altered in comparison to littermate control mice. For this purpose, we double-immunostained retina sections obtained from mice with different RIBEYE genotypes (wildtype: RBE^WT/WT; heterozygous knockin: RBE^WT/KI and homozygous knockin RBE^KI/KI) with various antibodies, including antibodies against opsin, to label photoreceptor outer segments (Figure 1A), SV2 (Figures 1B,C) and PSD95 (Figure 1A) to label retinal synapses, GFAP (Figure 1C) to label retinal Müller glia cells, and calbindin (Figure 1B) to mark horizontal cells. Using these markers, we did not observe gross alterations/differences in the retinal organization in mice with different RIBEYE genotypes (RBE^WT/WT; RBE^WT/KI, and RBE^KI/KI), arguing that the RBE^KI allele does not grossly affect the retinal organization. Similarly, the retinas of RIBEYE knockout mice, in which the RIBEYE protein was completely deleted, did not show gross alterations in retinal morphology and organization (Maxeiner et al., 2016).

Next, we checked for the expression of the RIBEYE B-domain in the RBE^KI mice (Figure 2) using a well-characterized mouse monoclonal antibody (clone 2D9) against the RIBEYE B-domain (Dembla et al., 2018). Both in wild-type RBE^WT/WT mice and in heterozygous RBE^WT/KI littermate mice, we observed strong punctate, ribbon-typical RIBEYE B-domain immunosignals in both synaptic layers (OPL and IPL) of the retina (Figure 2). The RIBEYE B-domain immunosignal was completely absent in the OPL and IPL of homozygous RBE^KI/KI mice (Figure 2C1). The absence of RIBEYE B-domain immunosignals in the homozygous RBE^KI/KI was expected because the RIBEYE B-domain was replaced in the RIBEYE KI allele by GCaMP3. Particularly strong RIBEYE B-domain immunosignals were observed in the OPL (Figure 2) of wildtype RBE^WT/WT mice and heterozygous RBE^WT/KI littermate mice because synaptic ribbons in rod synapses of the OPL are particularly large (Moser et al., 2020). For quantitative analyses, we thus first focused on the OPL of the retina in which the photoreceptor ribbon synapses are located.

Quantification of immunosignals in the OPL revealed that the RIBEYE B-domain immunosignals in the OPL were weaker in heterozygous RBE^WT/KI mice compared to littermate wildtype RBE^WT/WT, and confirmed the complete absence of RIBEYE B-domain immuno-signals in RBE^KI/KI mice (Figure 2D). Of note, the GCaMP3 moiety in the RIBEYE KI allele does not generate any fluorescence signals in resin sections of the immunolabeled retinas as shown by control incubations (Figure 2E).

The number and density of RIBEYE B-domain positive puncta in the OPL were determined on RIBEYE B-domain immunolabeled sections at higher magnifications that allow clear identification of single puncta (Figures 3A–C). Quantitative analyses revealed that the number of RIBEYE B-domain-positive puncta in the OPL were indistinguishable between RBE^WT/WT and heterozygous RBE^WT/KI mice (Figure 3D), but were completely absent from the OPL of homozygous RBE^KI/KI mice as expected (Figure 3D).

The RIBEYE B-domain antibody 2D9 also detects CtBP2 that is ubiquitously expressed as a nuclear co-repressor (Dembla et al., 2018). In the 0.5 μm thin retina sections, we did not observe a nuclear staining, similar to other studies (Dembla et al., 2018; Kesharwani et al., 2021), most likely because the focal concentration of CtBP2 in the nucleus is too low.

Is the RIBEYE A-domain still assembled into ribbon-like structures in RBE^KI mice? To address this question, we stained retinal sections with a polyclonal antibody against RIBEYE A-domain (Maxeiner et al., 2016). As mentioned above, the RIBEYE A-domain is still present on the RBE KI allele, in contrast to the RIBEYE B-domain that has been replaced. In wildtype RBE^WT/WT mice and heterozygous RBE^WT/KI littermate mice, we observed a strong punctate ribbon-typical RIBEYE A-domain immunosignal in both synaptic layers of the retina (OPL and IPL) (Figures 4A–C). Surprisingly, the RIBEYE A-domain signal was completely absent in the OPL of RBE^KI/KI mice (Figures 4C,D). Immunolabeling with an anti-SV2 antibody served as a control (Figure 4), demonstrating that synapses were still present as the SV2 immunosignals were unchanged in all genotypes (Figure 4).

The results obtained with the rabbit polyclonal antibody against the RIBEYE A-domain (Figures 4A–C) were confirmed with a newly generated mouse monoclonal RIBEYE A-domain antibody, clone 6F4 (Figure 5). The immunostaining pattern obtained with this antibody was indistinguishable from that produced with the polyclonal RIBEYE A-domain antibody, validating the absence of RIBEYE A-domain-positive puncta in the OPL and IPL of RBE^KI/KI mice (Figure 5C). In contrast, the RIBEYE A-domain immunostaining in the OPL and IPL of heterozygous RBE^WT/KI and wildtype RBE^WT/WT mice appeared unchanged. The newly generated mouse monoclonal RIBEYE A-domain antibody 6F4 is directed against a peptide stretch in the A-domain of RIBEYE (amino acid 83–amino acid 211 of mouse RIBEYE). The precise binding site of the antibody was determined by overlapping peptide dot blot experiments (Figure 5F). The specificity of the 6F4 antibody for RIBEYE was further demonstrated by Western blotting and immunofluorescence experiments on RIBEYE knockout mouse retina samples (Figures 5D,G,H). For the Western blotting experiments, the RIBEYE B-domain 2D9 monoclonal antibody was used for comparison. Both antibodies detected the full-length RIBEYE band at ≈120 kDa in control retinas, but not in retinas obtained from RIBEYE knockout mice (Figures 5D1,E1). The RIBEYE B-domain antibody 2D9 also detected CtBP2 at ≈50kDa in control and RIBEYE knockout retinas (Figure 5E1). CtBP2 is largely unaffected by the RIBEYE deletion in the RIBEYE knockout mice, as previously demonstrated (Dembla et al., 2018). The specificity of the RIBEYE A-domain antibody 6F4 was further confirmed by post-embedding immunogold labeling (Figure 5I). 6F4 antibody labeled synaptic ribbons in the control retina (Figure 5I1), but not in photoreceptor synapses of RIBEYE knockout retinas (Figure 5I4).

RIBEYE A-domain puncta densities in the OPL were determined using RIBEYE A-domain immunolabeled sections at higher magnifications that allow clear identification of single, immunolabeled puncta (Figures 6A–C,E), as described above for RIBEYE B-domain puncta density quantifications. Quantitative analyses revealed that the number of RIBEYE A-domain puncta in the OPL were indistinguishable between RBE^WT/WT and heterozygous RBE^WT/KI mice (Figure 6D). In contrast, RIBEYE A-domain puncta were completely absent from the OPL of RBE^KI/KI mice (Figure 6D). The results obtained with the rabbit polyclonal antibody against RIBEYE A-domain (Figures 6A–C) were confirmed by immunolabeling experiments with the monoclonal RIBEYE A-domain antibody 6F4 that also demonstrated the absence of RIBEYE A-domain puncta in the OPL (Figure 6E ).

These immunolabeling results, i.e., the absence of RIBEYE A-/RIBEYE B-immunolabeled puncta in the OPL of RBE^KI/KI mice, that we consistently obtained with anti-RIBEYE B-domain (Figures 2, 3) and anti-RIBEYE A-domain antibodies (6F4 and tau; Figures 4–6) indicate that no structures resembling synaptic ribbons are formed in photoreceptor synapses of RBE^KI/KI mice. To further test this conclusion, we performed ultrastructural analyses on rod photoreceptor synapses of the retinas from littermate RBE^WT/WT, RBE^WT/KI, and RBE^KI/KI mice by transmission electron microscopy (TEM; Figure 7).

TEM confirmed the absence of synaptic ribbons in photoreceptor synapses of RBE^KI/KI mice (Figures 7C1,C2), whereas synaptic ribbons were clearly present in RBE^WT/WT mice (Figures 7A1,A2) and heterozygous RBE^WT/KI (Figures 7B1,B2) (for quantification, see Figure 7D). Also, E-PTA-stained samples confirmed the absence of synaptic ribbons from the active zone of rod synapses in RBE^KI/KI mice (Figure 7C3). Synaptic ribbons were clearly present in E-PTA-stained rod photoreceptor synapses of RBE^WT/WT mice and heterozygous RBE^WT/KI mice (Figures 7A3,B3). From these experiments, we conclude that the RIBEYE B-domain is essential for the formation of synaptic ribbons in photoreceptor synapses.

As shown above, RIBEYE A- and B-domain immunosignals in the OPL were consistently weaker in heterozygous RBE^WT/KI mice in comparison to wild-type RBE^WT/WT mice (Figures 2, 4), whereas the ribbon density in the OPL was identical (Figures 3, 6). In order to identify possible reasons for the decreased RIBEYE immunosignals in the OPL of RBE^WT/KI mice in comparison to RBE^WT/WT mice, we measured the length of individual rod synaptic ribbons by 3D SR-SIM, as previously described (Dembla et al., 2020; Mukherjee et al., 2020; Kesharwani et al., 2021). For contour length measurements by 3D SR-SIM, rod synaptic ribbons were immunolabeled with either anti-RIBEYE B-domain (Figure 8) or anti-RIBEYE A-domain antibodies (Figure 9). By applying quantitative 3D SR-SIM analyses, we found that the contour length of synaptic ribbons was smaller in heterozygous RBE^WT/KI compared to RBE^WT/WT mice. This was consistently observed with both antibodies against RIBEYE (Figures 8, 9). This finding suggests that a decrease in the gene dosage of functional RIBEYE causes a decrease in the ribbon size, raising the exciting possibility that the ribbon size depends, among others, on the RIBEYE concentration.

In order to further characterize and confirm these light-microscopical alterations of synaptic ribbon size in photoreceptor synapses of RBE^WT/KI mice, we performed ultrastructural analyses by using TEM. We determined the height of synaptic ribbons in cross-sections of rod photoreceptor synapses of RBE^WT/WT mice and RBE^WT/KI mice. The height of cross-sectioned ribbons was measured from its anchorage site in the active zone to the free cytosolic end. For the determination of ribbon height, we applied the same criteria as previously defined (Kesharwani et al., 2021). Only rod photoreceptor synaptic ribbons were included that were anchored to a clearly visible active zone and were opposed by clearly visible postsynaptic triads. This procedure was applied to exclude those tangentially sectioned ribbons (sectioned parallel to the active zone) that were erroneously included in the analyses. We found that the ribbon height in rod photoreceptor synapses of heterozygous RBE^WT/KI mice was significantly lower (≈28%) than in RBE^WT/WT mice (Figures 10A,B; for quantification, Figure 10C).

The previous experiments consistently showed the absence of RIBEYE immunosignals in the OPL of RBE^KI/KI mice (Figures 2–6). As mentioned above, the OPL of the mouse retina contains mostly rod synapses, but also the less abundant cone synapses. In order to exclude the possibility that cone synapses might have escaped our attention in the immunolabeling experiments, we directly visualized cone terminals in the OPL by staining with fluorescent PNA lectin (Grabner et al., 2015) and analyzed with anti-RIBEYE 6F4 whether the cone terminals show RIBEYE immunosignals (Figures 11A1,A2). Despite the presence of PNA-positive cone terminals, no RIBEYE signals were observed in the OPL of RBE^KI/KI mice (Figure 11A2). The OPL of RBE^WT/WT mice served as a positive control in these experiments (Figure 11A1). Ribbons could be readily detected in the cone terminals of RBE^WT/WT mice (Figure 11A1) but not in the cone terminals of RBE^KI/KI mice (Figure 11A2) indicating the absence of synaptic ribbons also in cone synapses, similar to rod synapses.

To complement these analyses in the OPL, we analyzed synaptic ribbons in the inner plexiform layer (IPL) of the retinas from RBE^WT/WT, RBE^WT/KI, and RBE^KI/KI mice. For this purpose, we made use of the immunolabeling experiments that were shown in Figures 2, 4. In the IPL, 15 different types of retinal bipolar cells (Shekhar et al., 2016) form a morphologically and functionally heterogeneous population of synapses (Moser et al., 2020) that mostly contain synaptic ribbons at their active zone (Okawa et al., 2019). Synaptic ribbons are typically smaller in the IPL than in the OPL and exhibit different shapes, ranging from bar-shaped to ovoid (Moser et al., 2020). In agreement with the findings in the OPL, we observed a complete lack of both RIBEYE A-domain and RIBEYE B-domain immunosignals in the IPL of RBE^KI/KI mice in comparison to RBE^WT/KI and RBE^WT/WT mice (Figures 2, 4, 11B–E). We detected no RIBEYE fluorescence signals in the IPL of RBE^KI/KI mice (Figure 11) as expected. Electron microscopy confirmed the absence of synaptic ribbons in retinal bipolar cells, as shown by exemplary representative EM images of rod bipolar cells in Figures 11F–G. In RBE^WT/KI mice, RIBEYE fluorescence signals were again less intense than in the RBE^WT/WT mice, and RIBEYE puncta were also slightly, but highly significantly reduced (≈9% RIBEYE B puncta, ≈12% for RIBEYE A puncta) in comparison to homozygous RBE^WT/WT littermate control mice (for quantification, Figures 11B–E).

In order to determine whether deletion of the RIBEYE B-domain in RBE^KI mice has an impact on the expression levels of RIBEYE proteins, we performed Western blotting analyses on retinal lysates obtained from RBE^WT/WT, RBE^WT/KI, and RBE^KI/KI mice with the indicated primary antibodies using the Li-Cor system (Figure 12). As expected, we did not observe any RIBEYE band at ≈120 kDa with the anti-RIBEYE B-domain antibody in RBE^KI/KI mice (Figure 12B; for quantification, Figure 12F) because the RIBEYE B-domain is absent in RBE^KI/KI mice. In contrast, the RIBEYE band was clearly present in the lysates from littermate RBE^WT/WT and RBE^WT/KI mice (Figure 12B, arrowhead). With the anti-RIBEYE A-domain antibody, a RIBEYE A-domain-positive immunoblotting band was clearly present in retinal lysates from RBE^KI/KI mice, although the amount was less than in RBE^WT/WT mice and RBE^WT/KI mice (Figure 12C arrowhead; for quantification, Figure 12G). Antibodies against GFP were used to detect the expression of RIBEYE-GCaMP3 expression construct (Figure 12A, arrowhead). For normalization, the intensity of the respective bands obtained in the Li-Cor system was normalized to actin that served as a loading control. Importantly, the immunoblotting experiments revealed that the RIBEYE A-domain/GCaMP3 fusion protein was clearly produced in the RBE^KI/KI mice, but that the levels of the fusion protein were ≈31% of the wild-type RIBEYE levels (Figures 12C,G1,G2). These results suggest that the RIBEYE A-domain/GCaMP3 fusion protein may be unstable, possibly because it cannot be assembled into synaptic ribbons in the absence of the B-domain. The absence of synaptic ribbons in RBE^KI/KI mice is thus not due to a lack of RIBEYE protein synthesis but caused by the deletion of the B-domain.

Finally, we also analyzed synaptic ribbons in inner and outer hair cells (IHC and OHCs) in the organ of Corti of the inner ear. Synaptic ribbons were clearly visible in the IHCs and OHCs of RBE^WT/WT mice but completely absent in the IHCs and OHCs of RBE^KI/KI mice (Figure 13). This was consistently observed with both antibodies against RIBEYE B-domain (Figures 13A,B) as well as with antibodies against RIBEYE A-domain (Figures 13C,D).

Thus, synaptic ribbons were consistently absent in retinal synapses, both in the OPL and IPL, as well as in ribbon synapses in the IHCs and OHCs of the organ of Corti. These data show that the RIBEYE B-domain is essential for the assembly of synaptic ribbons.

RIBEYE is the defining protein component of synaptic ribbons that is essential for the formation of synaptic ribbons, as most directly shown with RIBEYE knockout mice in which deletion of RIBEYE leads to a complete absence of synaptic ribbons (Maxeiner et al., 2016; Becker et al., 2018; Jean et al., 2018). Although the fundamental role of RIBEYE in synaptic ribbons is thus well documented, the contribution of its individual protein domains and the mechanism of its action in ribbon assembly remains unclear. RIBEYE consists of a unique, proline-rich N-terminal RIBEYE A-domain and a C-terminal B-domain that is identical to CtBP2, a ubiquitously expressed transcriptional co-repressor. In support of the role of RIBEYE as a central building block of synaptic ribbons, multiple binding sites were identified in RIBEYE that mediate assembly RIBEYE multimers. In particular, several interaction sites were found in the RIBEYE A-domain that enable binding of the A-domain to other A-domains and to B-domains, which may be involved in ribbon assembly (Magupalli et al., 2008). Moreover, the RIBEYE B-domain/CtBP2 was shown to also interact with other B-domains and assemble into oligomeric structures (Kumar et al., 2002; Balasubramanian et al., 2003; Nardini et al., 2003; Madison et al., 2013; Bellesis et al., 2018; Jecrois et al., 2021). It seems likely that the RIBEYE A-domain constitutes the core assembly domain of the synaptic ribbon (Schmitz et al., 2000), but the role of the RIBEYE B-domain with its NAD(H) binding site in synaptic ribbons remains unknown. Indeed, it is even unclear whether the B-domain is required for synaptic ribbon assembly, or performs a different function associated with synaptic ribbons. For example, it is conceivable that the RIBEYE A-domain solely mediates the assembly of synaptic ribbons, and that the B-domain performs a peripheral function in tethering vesicles to the ribbon, an exciting possibility in view of the current lack of information on how synaptic ribbons organize continuous neurotransmitter exocytosis.

In the present study, we tested the basic role of the RIBEYE B-domain in mice. Using a genetic approach, we found that the RIBEYE B-domain is essential for the assembly of synaptic ribbons. Analysis of RIBEYE knockin (RBE^KI) mice in which the B-domain of RIBEYE was replaced by an unrelated protein (the Ca²⁺-sensor GCaMP3) revealed a complete absence of synaptic ribbons even though the RIBEYE A-domain continues to be expressed as a GCaMP3 fusion protein. Thus, the RIBEYE A-domain is not sufficient for ribbon assembly, and the phenotype caused by the deletion of the B-domain is as severe as that of the complete RIBEYE deletion (Maxeiner et al., 2016). Moreover, our data confirm that RIBEYE functions universally in all synaptic ribbons as previously suggested (Schmitz et al., 2000; Maxeiner et al., 2016; Becker et al., 2018; Jean et al., 2018), since, in RBE^KI/KI mice, synaptic ribbons were completely absent not only in the retina but also in the inner and outer hair cells of the organ of Corti (Figure 13).

Interestingly, we show that one genomic copy of RIBEYE B-domain is largely sufficient to enable the assembly of synaptic ribbons, but that the ribbons in heterozygous RIBEYE^WT/KI mice are smaller. In photoreceptor synapses, the number of ribbons (ribbon density) were unchanged in the heterozygous RIBEYE^WT/KI mice but decreased in size in comparison to RIBEYE^WT/WT mice (as judged by quantitative 3D SR-SIM and EM measurements; Figures 8–10). This decrease in ribbon size in heterozygous RIBEYE^WT/KI could be explained by two alternative hypotheses. It is possible that the decreased expression of RIBEYE protein in the retina of RIBEYE^WT/KI mice in comparison to RIBEYE^WT/WT mice translates into a difference in ribbon size. This would imply that the ribbon size is determined by the RIBEYE concentration similar to a mass-action-law. Alternatively, it is possible that trace amounts of the RIBEYE A-domain/GCaMP3 fusion protein are incorporated into the ribbons, and that they hinder the full assembly of the ribbons. However, the fact that smaller synaptic ribbons were also reported in heterozygous RIBEYE knockout mice (RBE^WT/KO) compared to RIBEYE WT mice (RBE^WT/WT) (Jean et al., 2018) strongly supports the first hypothesis. In further support of this hypothesis, moreover, transgenic overexpression of RIBEYE protein in zebrafish results in particularly large, oversized synaptic ribbons (Sheets et al., 2011, 2017). In the IPL of the inner retina, the fluorescence intensity of RIBEYE immunosignals was decreased in RIBEYE^WT/KI mice in comparison to RIBEYE^WT/WT mice. In the IPL, the number of synaptic ribbons/per IPL area was smaller in heterozygous RIBEYE^WT/KI mice in comparison to RIBEYE^WT/WT mice. This is probably based on the fact that synaptic ribbons in bipolar terminals in the IPL are typically smaller in size than in rod photoreceptor terminals (Moser et al., 2020). A further reduction in size in heterozygous RIBEYE^WT/KI mice thus can more easily result in their complete disappearance in bipolar cell synapses than in rod photoreceptor synapses in which synaptic ribbons are bigger.

In summary, our results demonstrate that the RIBEYE B-domain is essential for the assembly of synaptic ribbons and that the RIBEYE A-domain alone cannot build synaptic ribbons in the synaptic terminals of the retina and in sensory hair cells of the inner ear. How the RIBEYE B-domain promotes ribbon assembly needs to be further analyzed, but it seems likely that it acts as more than a structural component. Recent biochemical and structural analyses on CtBP1 and CtBP2, that is largely identical to the RIBEYE B-domain, can give important hints (Nardini et al., 2009; Madison et al., 2013; Bellesis et al., 2018; Jecrois et al., 2021). These data showed that CtBP2 forms tetrameric complexes via dimeric intermediates (Madison et al., 2013; Nichols et al., 2021). The assembly of these complexes is enhanced by NAD(H) (Madison et al., 2013; Jecrois et al., 2021; Nichols et al., 2021). Interestingly, NAD(H) has been shown to influence the size of synaptic ribbons in sensory hair cells (Wong et al., 2019; Okur et al., 2020). The size of auditory hair cell ribbon synapses also strongly increases with aging in C57BL/6J mice (Peineau et al., 2021). Based on the CtBP1/CtBP2 structural data (Balasubramanian et al., 2003; Madison et al., 2013; Jecrois et al., 2021; Nichols et al., 2021), the RIBEYE B-domain can be expected to also assemble into tetrameric complexes that are controlled by NAD. These oligomeric RIBEYE B-complexes could link different RIBEYE units to each other during ribbon assembly and/or mediate stability to the synaptic ribbon. The functional properties of ribbon synapses in RBE^KI/KI mice and the physiological consequences of RIBEYE B-domain deletion in the RBE^KI/KI mice for vision and hearing need to be analyzed by future analyses.

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author/s.

The animal study was reviewed and approved by Landesamt für Verbraucherschutz; Geschäftsbereich 3; 66115 Saarbrücken, Germany; GB 3-2.4.1.1-K110/180-07.

SS performed all experiments and analyses shown in Figures 1–13 and wrote the article together with FS and TS. KS performed embeddings and provided help with experiments and statistical procedures. RK introduced SS to confocal microscopy and provided help with analyses. MJ generated the peptide arrays and provided important advice. SM generated knockin and knockout mice and provided help with analyses. TS designed knockin strategy, organized KI and KO mouse generation and provided essential support. FS designed and supervised the study and wrote the article together with SS and TS. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.