Role of Ribeye PXDLS/T-binding cleft in normal synaptic ribbon function

Abstract

Non-spiking sensory hair cells of the auditory and vestibular systems encode a dynamic range of graded signals with high fidelity by vesicle exocytosis at ribbon synapses. Ribeye, the most abundant protein in the synaptic ribbon, is composed of a unique A domain specific for ribbons and a B-domain nearly identical to the transcriptional corepressor CtBP2. CTBP2 and the B-domain of Ribeye contain a surface cleft that binds to proteins harboring a PXDLS/T peptide motif. Little is known about the importance of this binding site in synaptic function. Piccolo has a well-conserved PVDLT motif and we find that overexpressed Ribeye exhibits striking co-localization with Piccolo in INS-cells, while two separate mutants containing mutations in PXDLS/T-binding region, fail to co-localize with Piccolo. Similarly, co-transfected Ribeye and a piccolo fragment containing the PVDLT region co-localize in HEK cells. Expression of wild-type Ribeye-YFP in zebrafish neuromast hair cells returns electron densities to ribbon structures and mostly rescued normal synaptic transmission and morphological phenotypes in a mutant zebrafish lacking most Ribeye. By contrast, Ribeye-YFP harboring a mutation in the PXDLS/T-binding cleft resulted in ectopic electron dense aggregates that did not collect vesicles and the persistence of ribbons lacking electron densities. Furthermore, overexpression failed to return capacitance responses to normal levels. These results point toward a role for the PXDLS/T-binding cleft in the recruitment of Ribeye to ribbons and in normal synaptic function.

Figure 1.. Colocalization of overexpressed Ribeye with presynaptic proteins in INS-1 cells.

INS-1 cells were transfected with rat Ribeye-GFP and then immuno-stained for Ribeye and several presynaptic proteins. (A-D) show example images from co-staining for Ribeye and Piccolo (A), RIM2 (B), CaV1.3 (C) and Bassoon (D). For each set of images, magenta denotes the presynaptic protein of interest, green indicates GFP and the right panel merges both colors. Note the striking degree of colocalization between spots of Piccolo and GFP in A; scale bar: 5µm applies to all images. (E) Colocalization was quantified by determining the ICQ value. The ICQ varies from −0.5 to 0.5, with the higher numbers indicating better correlation between the two channels. For each protein pair, an ICQ value was determined for Piccolo n=17, RIM2 n=14, CaV1.3 n=4 and Bassoon n=6 transfected cells images and averaged. *P<0.05.

Figure 2.. Ribeye-Piccolo colocalization requires intact PXDLS/T-binding cleft.

(A-C): Examples of cells transfected with Ribeye-EGFP (A) or Ribeye-EGFP with two different mutations in the PXDLS/T binding cleft (B-C). Green denotes EGFP, whereas Magenta denotes Piccolo localization. D, Degree of correlation was calculated using the ICQ quantification. Wild type Ribeye n=18, D583-6A n=12, V615R n=20 cells. Scale bar: 5 µm. *P<0.05.

Figure 3.. Cleft-filled Ribeye mutant fails to localize to Piccolo in INS-1 cells.

INS-1 cells were transfected with construct for expression of either rat Ribeye-HA (A) or a ‘cleft-filled’ rat Ribeye (B), which has the E1A sequence concatenated to its C-terminus. E1A binds to the B-domain of Ribeye and thus is expected to compete with Piccolo for binding at the PXDLS/T binding cleft. Ribeye constructs are shown in magenta and piccolo staining is shown in green. Scale bar is 5 µm.

Figure 4.. Co-expression of Ribeye and Piccolo (2622–2937) in HEK cells shows striking colocalization that is disrupted in cleft-filled mutant.

HEK cells were co-transfected with a fragment of Piccolo containing the PVDLT sequence (Piccolo(2622–2937)) and either EGFP (A) or variants of Ribeye (B-D). Piccolo(2622–2937) exhibited striking colocalization with Ribeye-EGFP (B) and Ribeye-HA (C), but not the cleft filled Ribeye (D; rRE-E1AC-HA) or GFP. Scale bar: 5 µm

Figure 5.. Piccolo PVDLT containing peptide labels ribbons in bipolar cells.

Goldfish retinal bipolar cells were loaded with a fluorescent Ribbon-binding peptide (RBP; magenta) and a 14 amino acid peptide derived from Piccolo that encompasses the PVDLT motif (PCLO-14; green) and imaged using TIRF microscopy. (A) bright-field image of bipolar cell terminal. (B) Merged TIRF image of PCLO-14 (green) and RBP (magenta) peptides. Same cell as in A, C, D. Scale bar: 10 µm in B.

Figure 6.. Ribeye-Piccolo colocalization requires intact PXDLS/T-binding cleft.

A-B: ribeye (a(∆10)/b(∆7) zebrafish were injected with plasmids for either the short isoform of Ribeye a concatenated to YFP (Ribeye as-YFP) or to the short isoform of Ribeye a with a mutation to its PXDLS/T-binding cleft (Ribeye as(V338R)-YFP) co-stained with the post-synaptic marker pan-MAGUK. (A) Expression of WT-Ribeye in the double-homozygous mutant zebrafish detected by ctbp2 antibody shows a typical colocalization to postsynaptic density marker pan-MAGUK. (C) and (D) Enlargement of ROIs shown in A; (E) and (F) magnified image of ROI in (B). (G) To determine synaptic localization of Ribeye spots, the distance of each Ribeye spot was measured to the nearest pan-MAGUK spot. (H) Circles denote distances between each individual Ribeye spot and nearest pan-MAGUK spot, bar graph represents the mean across spots. (I) same as in H, but showing distributions of averages across neuromasts. Error bars indicate SEM. ***P<0.001, **P<0.01.

Figure 7.. Ribeye(a)s-YFP overexpression partially rescues the lost synaptic density in ribbons of neuromast hair cells in ribeye(a(∆10)/b(∆7) mutants.

(A-B) Electron micrograph showing typical examples of ghost ribbon (denoted by *) in neuromast hair cells in ribeye(a(∆10)/b(∆7)) zebrafish mutants. (C-F) Examples of electron micrographs from Ribeye(a)s-YFP over-expressing hair cells in the in ribeye(a(∆10)/b(∆7)) zebrafish mutants. Note the electron densities in the structures with diverse morphologies (arrows). Most densities were spherical, but were smaller than densities in WT ribbons, but some showed bar-like laminar structures (e.g. lower right density in E), reminiscent of retinal ribbons. All images taken from 5 days post-fertilization fish. Scale bar is 500 nm. Note that despite partial return of density to ribbons, many ribbons still fail to localize to synaptic locations.

Figure 8.. Ribeye as-(V338R)-YFP overexpression fails to rescue the lost synaptic density in ribbons of neuromast hair cells in ribeye(a(∆10)/b(∆7) mutants.

A-E: Examples of electron micrographs from WT Ribeye as-(V338R)-YFP over-expressing hair cells in the in ribeye(a(∆10)/b(∆7)) zebrafish mutants. A) low magnification image showing that electron densities scattered throughout neuromast of Ribeye as-(V338R)-YFP expressing hair cells. (B, C) Higher magnification images showing that unlike Ribeye(a)s-YFP expressing cells (figure 7), electron densities are not surrounded by vesicle-containing ribbon-ghosts. (D) Example of ribbon-ghosts lacking a density (denoted by *) in Ribeye as-(V338R)-YFP expressing cell. (E) Example of giant electron dense structures occasionally found in Ribeye (a)s-(V338R)-YFP expressing cells. (F) Areas of ribbons or ghost-ribbons in WT, ribeye(a(∆10)/b(∆7)) and ribeye(a(∆10)/b(∆7)) overexpressing either Ribeye(a)s-YFP or Ribeye (a)s-(V338R)-YFP. Note that Ribeye (a)s-YFP causes an increase in ribbon size, but ribbons still remain smaller than WT ribbons. *P<0.05, **P<0.01, ***P<0.001.

Figure 9.. Measurements of electron densities.

(A-C) Examples of electron micrographs of WT ribbons in hair cells of zebrafish neuromasts. (D) Area of electron densities in WT animals and in ribeye(a(∆10)/b(∆7)) mutants overexpressing either Ribeye(a)s-YFP or Ribeye(a)s-(V338R)-YFP. Note that electron densities in Ribeye(a)s-(V338R)-YFP were outside of ribbon structures.

Figure 10.. Lateral line hair cells overexpressing Ribeye(a)s-YFP, but not those expressing Ribeye(a)s-(V338R)-YFP reduce calcium currents and capacitance responses.

Neuromast hair cells from 4–9 dpf zebrafish were whole-cell voltage clamped and capacitance measurements were performed to track exocytosis. Hair cells were subjected to 3 s long depolarization to −10 mV and the difference after the depolarization relative to the baseline before stimulation was measured. Zebrafish embryos were injected at the one cell stage with a construct encoding for either Ribeye-as-YFP or Ribeye(a)s-(V338R)-YFP. Fluorescent cells containing the transgenes were compared against non-fluorescent cells in the same fish as controls. (A) A shows the average capacitance recordings of hair cells expressing Ribeye-as-YFP and Ribeye(a)s (V338R)-YFP in zebrafish(a(∆10)/b(∆7). (B) Hair cells were depolarized to −20 mV and leak-subtracted currents were measured for same animals as in A. Bar graphs show mean +/− SEM. *P<0.05.