Munc18-1 is crucial for photoreceptor function, survival and regulation of syntaxin-3 localization and expression

Abstract

The function of the SNARE complex regulator, Munc18-1, in photoreceptor cells is unknown. Here, we found that removing Munc18-1 from photoreceptors results in major degeneration starting at P14. In the absence of Munc18-1, before major photoreceptor degeneration, functional and synaptic impairments were present, indicating a critical function of Munc18-1. Furthermore, Munc18-1 played a critical role in expression and localization of syntaxin-3. The syntaxin-3 protein level is dramatically reduced in the soma and plasma membrane of photoreceptors without Munc18-1. At the photoreceptor synapses, the colocalization of syntaxin-3 and its SNARE partner, SNAP-25, was reduced, potentially suggesting an altered syntaxin-3 synaptic localization. In the Munc18-1-deficient photoreceptors, immature synapses and outer segment lesions were found. Taken together, these findings provide evidence that Munc18-1 is important for maintaining sufficient syntaxin-3 expression in the cell body and synapses of photoreceptors. The lack of Munc18-1, combined with poor syntaxin-3 expression, contributes to photoreceptor functional impairment and degeneration.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-57827-1.

Introduction

The retina is the light sensitive tissue at the back of the eye and is made up of 5 types of neurons. These cells stratify into three distinct nuclear layers that are separated by two synaptic plexiform layers for cell-to-cell communication^1,2. Visual signals are transduced by photoreceptor cells located in the outer nuclear layer (ONL), passed onto bipolar cells in the inner nuclear layer (INL), and then onto ganglion cells³. Retinal synapses are highly ordered, complete their development postnatally, and form well-defined connections, making the retina an exceptional tool to study synapse development, refinement, and maintenance.

In the outer retina, photoreceptor cells adjust glutamate release to convey light information to their postsynaptic bipolar cells using special synapses termed ribbon synapses. Synaptic vesicles are tethered to this proteinaceous ribbon structure in preparation for exocytosis. The SNARE complex, composed of isoforms of syntaxin, synaptosomal associated protein (SNAP), and synaptobrevin/vesicle-associated membrane protein (VAMP), is the main machinery executing membrane fusion and exocytosis for neurotransmitter release at conventional synapses^4–6. Despite previous suggestions that exocytosis at ribbon synapses may occur through SNARE-independent mechanisms⁷, it is now evident that exocytosis at photoreceptor and inner hair cell ribbon synapses is SNARE dependent, albeit with alternate SNARE isoforms⁸. Specifically, we and others established that syntaxin-3 and SNAP-25 are the obligate target SNARE (t-SNARE) proteins utilized for synaptic transmission as well as outer segment dynamics in photoreceptor cells^9–11. However, how SNARE-mediated exocytosis is regulated at photoreceptor cell ribbon synapses remains unknown.

Complementary to SNARE proteins, Sec1/Munc18-like (SM) proteins help to orchestrate membrane fusion in conventional synapses. Munc18-1, also known as STXBP1, was originally characterized as a syntaxin-1 binding protein¹². Munc18-1 is required for neuron survival and neurotransmitter release as shown by Munc18-1 null mice¹³ and Munc18-1 null neuronal culture^14,15. However, because Munc18-1 deficiency compromises neuron survival, how it organizes SNARE proteins in presynaptic nerve terminals is poorly understood. For example, Munc18-1 appears to be required for syntaxin-1 expression in neurons, where in the absence of Munc18-1, syntaxin-1 levels are reduced by 70%, although the remaining syntaxin-1 is reported to be correctly targeted to synapses¹⁶. On the other hand, a perturbation in plasmalemmal syntaxin-1 localization has been demonstrated in Munc18-1 knockdown or Munc18-1/-2 double knockdown PC12 cells^17,18. Similar changes in syntaxin-1 distribution have also been reported in soma of neurons that are differentiated from patient-derived Munc18-1 heterozygous mutant iPSCs¹⁹. The rapid death of neurons makes investigating Munc18 function and protein interactions in primary neurons challenging, but surrogate systems such as PC12 and iPSCs lack true synapses. Thus, how Munc18-1 regulates syntaxin transport to presynaptic nerve terminals remains elusive. Further, the consequences of disrupting Munc18-1 and syntaxin interaction on other SNARE proteins or SNARE complex localization are unclear.

A heterozygous mutation (H16R) in Munc18-1 is known to be associated with congenital nystagmus in humans²⁰. Nevertheless, the exact roles and functions of Munc18-1 in photoreceptor cells and vision remain unknown. Synaptic vesicle exocytosis must be carefully and precisely regulated in photoreceptors, as neurotransmitter release is mainly tonic in the dark with occasional phasic release coding stimulus changes²¹. Light stimulation hyperpolarizes photoreceptor cells, leading to reduced glutamate release. Sustaining the massive glutamate release in photoreceptor cells under dark conditions, along with mediating rapid modulation of vesicle release in response to light relies on the activity of key regulatory proteins. Therefore, it is crucial to identify Munc18-1’s role in SNARE protein targeting in photoreceptor ribbon synapses.

To address these questions, we generated photoreceptor-specific Munc18-1 conditional knockout (cKO) mice, wherein CRX-driven cre-recombinase induces the removal of Munc18-1 from photoreceptors starting on embryonic day E12.5²². Using behavioral and functional vision tests, combined with histology and ultrastructural analyses, we examined the consequences of Munc18-1 loss in photoreceptor cells. Despite significant impairment to photoreceptor cell functioning by postnatal day 14 (P14), remarkably, the majority of Munc18-1 cKO photoreceptors survived until P14, providing an opportunity to investigate the interplay between Munc18-1 and other SNARE proteins in vivo. Here, we addressed two key questions: (i) whether syntaxin-3 expression and localization are dependent on Munc18-1, and (ii) whether other SNAREs are impacted following the removal of Munc18-1. Our results demonstrated that Munc18-1 is essential for photoreceptor cell survival and functions, and that Munc18-1 is required for the proper expression and localization of syntaxin-3 in synaptic terminals.

Results

Munc18-1 removal from photoreceptors leads to retina thinning and loss of visual function

To investigate the roles of Munc18-1 in photoreceptor neurons, we specifically removed Munc18-1 from photoreceptor cells and a subset of bipolar neurons by crossing Munc18-1 flox mice²³ with CRX-cre mice²². Because CRX-Cre primarily targets photoreceptor cells, we refer to this mouse line as a photoreceptor-specific Munc18-1 knockout (Munc18-1 cKO). We started with in vivo retinal imaging and behavioral tests to identify gross morphological and functional changes at P22 when photoreceptor cell maturation is complete. Optical coherence tomography (OCT) imaging showed that Munc18-1 cKO mice had reduced retinal thickness by P22 (Supplemental Fig. 1A). Electroretinogram (ERGs) analysis under dark-adapted (scotopic) and light-adapted (photopic) conditions showed that Munc18-1 cKO photoreceptor cells did not hyperpolarize in response to light (no A-wave) and also failed to induce a downstream B-wave (Supplemental Fig. 1B). Finally, adult Munc18-1 cKO mice had a profound reduction in visual acuity, as measured by the optokinetic tracking response when compared to the controls (Supplemental Fig. 1C). Therefore, we next moved to analyze the section histology of Munc18-1 cKO and littermate eyes to corroborate the changes we saw by OCT. We first checked Munc18-1 control and cKO mice at the same timepoint when electroretinograms and OCT scans were performed. Histological analysis of the retina of Munc18-1 cKO mice at P22 revealed a severe loss in total retina thickness in Munc18-1 mice, as almost zero photoreceptor cells remained by this age (Supplemental Fig. 1D), confirming the retina thickness decrease we observed by OCT (Supplemental Fig. 1A).

Munc18-1 deficient photoreceptors degenerate between P14 and P22

To understand why photoreceptor cells do not survive to P22 in Munc18-1 cKO mice, we examined photoreceptor cells at earlier timepoints (P9, P12, and P14) to further investigate structural and functional changes. We first confirmed the successful removal of Munc18-1 using a polyclonal Munc18-1 antibody. In control sections, Munc18-1 was abundantly present in photoreceptor cells (Fig. 1A, left, control panels). Using a different antibody, we observed strong Munc18-1 staining in the outer plexiform layer (OPL), where synapses are found, and outer nuclear layer (ONL) (Fig. 1A, right, control panels). Although weaker, staining of Munc18-1 was also present in photoreceptor outer and inner segments (OS and IS) (Fig. 1A). We confirmed the specificity of BD mouse anti-Munc18-1 antibody by having a secondary antibody only control, where the ONL and OPL were not stained (Supplemental Fig. 2A). Western blot using the BD Munc18-1 antibody detected strong signal in the isolated retina samples, but not the retinal pigment epithelium (RPE) (Supplemental Fig. 2B,C). Together, this indicated that the staining of Munc18-1 is strong in the photoreceptors, but not in the RPE.

Fig. 1.

Effective removal of Munc18-1 by CRX-cre leads to degeneration between postnatal day 14 and 22. (A) Left: Antibody staining for Munc18-1 (green) using rabbit polyclonal antibody against PSD95 (red) for presynaptic terminals in P9 control and P9 Munc18-1 retina. Munc18-1 is present throughout the retina in the outer nuclear layer (ONL), outer plexiform layer (OPL), inner nuclear layer (INL), inner plexiform layer (IPL), and ganglion cell layer (GCL). In control retina, a meshed/plasmalemmal pattern of Munc18-1 is observed surrounding photoreceptor soma. Munc18-1 fluorescence is lost in the photoreceptor and bipolar cells of Munc18-1 cKO mice. Retina region for all represented images is 0.5–1 mm away from optic nerve. All images are maximum intensity projections. Scale bar = 20 μm. Right: Antibody staining for Munc18-1 (green) using mouse monoclonal antibody and PSD95 (red) in P9 control and Munc18-1 cKO retina. A decrease in Munc18-1 fluorescence is signal is observed in Munc18-1 cKO mice. Retina region for all represented images is 0.5–1 mm away from optic nerve. All images are maximum intensity projections. Scale bar = 50 μm. (B) Antibody staining for Munc18-1 (green) using rabbit polyclonal antibody against PSD95 (red) for presynaptic terminals in P12 and P14 control and Munc18-1 cKO retina. Top: scale bar = 50 μm. Bottom: scale bar = 20 μm. All images are maximum intensity projections. (C) Quantification of outer nuclear layer (ONL) thickness in control, Munc18-1 cKO (green, squares) mice and SNAP-25 cKO (purple, circles) mice. Control mice for both lines are their respective flox/flox littermates, i.e. SNAP-25 flox/flox vs. SNAP-25 cKO and Munc18-1 flox/flox vs. Munc18-1 cKO. SNAP-25 cKO photoreceptors are rapidly lost between P9 and P14 while Munc18-1 cKO photoreceptors are rapidly lost between P14 and P22. Although Munc18-1 flox/flox control ONLs were thicker than Munc18-1 cKO ONLs, they were not statistically significant at any age before P14 (two-way ANOVA; P9, p = 0.9976; P12, p = 0.1992; P14, p = 0.5107). SNAP-25 exhibit significant degeneration after P9 (two-way ANOVA; P9, p = 0.7049; P12, p < 0.0001; P14, p < 0.0001). Error bar represents SEM. (D) Averaged scotopic (dark-adapted) and photopic (light-adapted) electroretinogram traces across 7 light intensities in P14 control and P14 Munc18-1 cKO mice (solid line). Arrow indicate stimulation onset, A indicates a-wave, B indicates b-wave. Munc18-1 cKO mice exhibit diminished light responses at all light intensities. Each data point represents one of eye each mouse; one eye was excluded from a control mouse due to eyelid deformity. Error bar represents SEM.

Presynaptic photoreceptor terminals were identified using the presynaptic marker PSD95. Munc18-1 was present in the PSD95 positive layer, indicating its presence at the synaptic OPL (Fig. 1A, control panels). The Munc18-1 signal in the ONL was present surrounding photoreceptor nuclei, creating a characteristic meshed/webbed pattern in the ONL. As photoreceptor nuclei are compactly packed in the ONL, Munc18-1 was likely present on the plasma membrane and within a thin cytosol region in photoreceptors. Munc18-1 was also strongly observed throughout the inner nuclear layer (INL), inner plexiform layer (IPL), and ganglion cell layer (GCL)/retinal nerve fiber layer. However, in Munc18-1 cKO mice, Munc18-1 fluorescence was largely absent surrounding photoreceptor soma in the ONL and the meshed staining pattern was lost (Fig. 1A, Munc18-1 cKO panels). In the OPL, Munc18-1 colocalization with PSD95 was mostly absent, confirming effective Munc18-1 knockout in photoreceptor synapses (Fig. 1A, Munc18-1 cKO panels). Munc18-1 was also decreased in the apical INL, indicating Munc18-1 removal from a subset of bipolar cells⁹. Horizontal cells were unaffected by our conditional CRX-cre induced knockout. We confirmed that Munc18-1 staining in the OPL of Munc18-1 cKO partially colocalized with calbindin, a marker of horizontal cells (Supplemental Fig. 3). Such staining suggests that the majority of Munc18-1 from photoreceptors was removed, with the remaining OPL contributed by various cell types in the inner nuclear layer. The inner retina was left intact as Munc18-1 remained abundantly expressed throughout the basal INL and IPL. Despite Munc18-1 being efficiently removed in P9 photoreceptor cells, we did not observe a significant decrease in photoreceptor layer thickness at P9 in Munc18-1 cKO mice when compared to age-matched controls (Fig. 1C). We did not observe significant photoreceptor loss leading up to P14 (Fig. 1C), and the majority of photoreceptor loss largely occurred between P14 and P22 in Munc18-1 cKO retinas.

While Munc18-1 cKO ultimately led to photoreceptor cell degeneration, photoreceptors were largely maintained at P14, we did not exclude that photoreceptors may already be unhealthy before degeneration. We took advantage of this and tested the ERG response of Munc18-1 cKO mice at P14 at eye opening, which is the earliest possible time to perform ERG (Fig. 1D). We found that despite the majority of photoreceptor cells being maintained in Munc18-1 cKO mice, their ERG response at P14 was strongly reduced (Fig. 1D). Our results suggested that Munc18-1 is necessary for light responsiveness due to the large reduction in photoreceptor A-wave. The diminished B-wave could be attributed to the overall depressed A wave or potentially decreased synaptic communication.

Altered syntaxin-3 expression and localization in the soma and outer segments of Munc18-1 deficient photoreceptor cells

It has previously been reported that in the absence or mutation of Munc18-1, syntaxin-1 is mislocalized in the soma of PC12 cells and iPSC-derived neurons^17–19. Therefore, we investigated the localization of syntaxin-3 in Munc18-1 cKO retinas. In the control retina, Munc18-1 and syntaxin-3 were both found throughout photoreceptor cells in their synapses, photoreceptor soma, and weakly in the outer and inner segments (Fig. 2A). Syntaxin-3 localization was highly consistent with previous findings^9,24,25. The ONL pattern of syntaxin-3 was similar to that of Munc18-1, in which syntaxin-3 surrounded photoreceptor nuclei with an apparent meshed staining pattern. Syntaxin-3 is a transmembrane protein, and therefore, the meshed pattern suggests an enriched localization in the photoreceptor plasma membrane. Strong colocalization of Munc18-1 and syntaxin-3 was present in the control sections, indicating Munc18-1 was likely positioned closely with the plasma membrane (Fig. 2A). However, in P12 Munc18-1 cKO retina sections, syntaxin-3 no longer exhibited the meshed appearance that was obvious in the control sections. We observed a strong decrease in meshed syntaxin-3 signal in Munc18-1 cKO sections except for small regions where cre-recombination was less effective, and Munc18-1 remained present (Fig. 2A).

Fig. 2.

Syntaxin-3 mislocalisation in Munc18-1 cKO mice. (A) Munc18-1 (green), syntaxin-3 (red), and RIBEYE (white) stain in P12 control and Munc18-1 cKO retina. RIBEYE is used as a structural protein to identify the ribbon synapse of photoreceptors. Munc18-1 and syntaxin-3 show strong colocalization in control retina and is found in photoreceptor outer segments, synapses, and in nuclear plasmalemmal. Both Munc18-1 and syntaxin-3 show meshed plasmalemmal staining pattern throughout the ONL. In Munc18-1 cKO retinas, regions where Munc18-1 is absence also have decreased plasmalemmal syntaxin-3 signal. Syntaxin-3 remains abundantly present in Munc18-1 cKO synapses due to the continued share localization with the synaptic protein RIBEYE. In Munc18-1 cKO sections, punctate aggregates of syntaxin-3 are also observed in photoreceptor inner and outer segments. All images are maximum intensity projections. Scale bar = 20 μm. (B) Magnified region from (A), in control retina, syntaxin-3 is present surrounding photoreceptor nuclei in a meshed pattern and in the photoreceptor inner/outer segments as well as photoreceptor terminals. In Munc18-1 cKO sections, the meshed pattern of syntaxin-3 is absent throughout the ONL, and instead perinuclear syntaxin-3 aggregates are observed in the ONL (yellow arrow heads). In control sections, syntaxin-3 is uniformly found through the ciliary outer segment structures of control photoreceptors. In Munc18-1 cKO retina, the ciliary staining pattern is lost, however, punctate syntaxin-3 aggregates can be observed in photoreceptor inner and outer segments (yellow arrows). Synaptic targeting of syntaxin-3 remains intact in Munc18-1 cKO as syntaxin-3 occupies the presynaptic OPL as identified by RIBEYE.

Concurrent with the disappearance of syntaxin-3 around photoreceptor nuclei, we observed aggregation of syntaxin-3 in the ONL cell bodies, inner/outer segment regions of Munc18-1 cKO photoreceptors, and in photoreceptor synaptic terminals (Fig. 2A,B). Circular puncta of syntaxin-3 aggregates were observed near photoreceptor nuclei, and in the ciliary segment region of Munc18-1 deficient photoreceptor cells (Fig. 2B). We used 3D reconstruction to visualize and quantify the syntaxin-3 puncta in Munc18-1 cKO retina sections (Supplemental Fig. 4), where we saw a significant increase in syntaxin-3 puncta in the ONL. Therefore, without Munc18-1, syntaxin-3 at the photoreceptor cell soma was reduced in expression and was not properly localized, resulting in a loss of the meshed pattern and formation of syntaxin-3 puncta.

Despite the decrease of meshed syntaxin-3 staining in Munc18-1 cKO, syntaxin-3 was expressed in photoreceptor synaptic terminals, as syntaxin-3 signal was enriched in the OPL with RIBEYE, a marker for the synaptic ribbon²⁶ (Fig. 2A,B). Enrichment of syntaxin-3 signal in the vicinity of RIBEYE in both control and Munc18-1 cKO synaptic terminals suggests that transport of syntaxin-3 to the photoreceptor synapse is largely preserved in the absence of Munc18-1. This result was consistent with the previous findings from embryonic day 18 (E18) brain slices of Munc18-1 null mice, where the remaining syntaxin-1 was targeted to synaptic regions¹³.

Changes in syntaxin-3 localization in the photoreceptor cells could be an artefact of severe degeneration. Therefore, we examined syntaxin-3 localization in SNAP-25 cKO mice, which also exhibit severe photoreceptor cell degeneration (Fig. 1C) ¹¹. Both t-SNARES syntaxin-3 and SNAP-25 colocalized in control sections as expected (Supplemental Fig. 5). In SNAP-25 cKO mice, syntaxin-3 continued to be present surrounding photoreceptor nuclei and in the synapses (Supplemental Fig. 5). Therefore, we concluded that the changes in the syntaxin-3 staining surrounding the soma of photoreceptors observed in Munc18-1 cKO retinas are not the consequence of severe degeneration, and rather a specific effect of Munc18-1 deletion.

Alternatively, accumulation of syntaxin-3 could represent dendritic sprouting from secondary neurons, such as horizontal cells and bipolar cells. Dendritic sprouting occurs with aging and synaptic mutants such as Bassoon mutant mice^27,28. However, calbindin staining in Munc18-1 cKO does not show processes extending from horizontal cells into ONL (Supplemental Fig. 3). Bipolar cells also showed no evidence of postsynaptic invagination into the ONL (Supplementary Fig. 6A, B). Dendritic sprouting in many cases form ectopic synapses in the ONL²⁷. Nonetheless, staining of presynaptic marker RIBEYE did not colocalize with syntaxin-3 puncta in ONL (Fig. 2A,B). Thus, although there could be a few postsynaptic invaginations immediately above the OPL, the majority of syntaxin-3 puncta reflect improper localization of syntaxin-3.

Removal of Munc18-1 impairs syntaxin-3 localization with respect to SNAP-25 within presynaptic terminals

Since the localization of syntaxin-3 was changed in the soma, inner segment and outer segment of Munc18-1 deficient photoreceptors, we next asked whether this change is specific to syntaxin-3, or if Munc18-1 removal affected the localization of the other t-SNARE, SNAP-25. Target-SNARE partners syntaxin-3 and SNAP-25 were both abundantly present in photoreceptor cells and showed strong colocalization in the control retina, as expected (Fig. 3A). We observed no differences in ONL or synaptic SNAP-25 expression pattern between the Munc18-1 cKO retina versus control retina. Similar to previous findings¹¹, SNAP-25 was abundantly found in the ONL around the nuclei, synapses, and throughout the inner retina (Fig. 3A). We concluded that SNAP-25 localization in photoreceptor cell bodies and synapses was unaffected by Munc18-1 deletion (Fig. 3A). However, we noted that some aggregates of SNAP-25 were present in the inner and outer segments of Munc18-1 cKO photoreceptors, which may be non-specifically caused by degeneration, similar to the syntaxin-3 accumulations observed above.

Fig. 3.

Reduction of colocalization between t-SNAREs Syntaxin-3 and SNAP-25 in Munc18-1 cKO. (A) Syntaxin-3, (green) and SNAP-25, (red) immunostaining in P12 control and Munc18-1 cKO retinas. Syntaxin-3 and Munc18-1 strongly colocalize in maximum intensity projections, (left) of control retina sections, (yellow). Syntaxin-3 loses plasmalemmal somatic locale in Munc18- cKO retinas while SNAP-25 signal remains unchanged in Munc18-1 cKO retina. Synaptic Syntaxin-3 and SNAP-25 no longer colocalize together in Munc18-1 cKO retinas, clearly visualized in magnifications of single z-planes presented on right. (B) Intensity profile of syntaxin-3, (green) and SNAP-25, (red) in P12 control and Munc18-1 cKO retinas. An overlap in syntaxin-3 and SNAP-25 intensity is observed in control retinas, consistent with the observed colocalization. However, the peaks of syntaxin-3 and SNAP-25 signal intensity are shifted and no longer overlap in Munc18-1 cKO retinas and is consistent with the loss of outer plexiform layer colocalization observed in, (A). Quantification is of 3 regions of 3 different sections, (n = 3), profile quantification was performed across N = 3 individual animals per group., (C) Representative 3D volume projections of the outer plexiform layer in P12 control and Munc18-1 cKO retina. A decrease in the volume of syntaxin-3 and SNAP-25 overlap is observe in Munc18-1 cKO retina. (D) Quantifications of syntaxin-3 and SNAP-25 volume overlap, t(16) = 3.862; p = 0.0014. (E) Quantifications of syntaxin-3 volume not colocalized with SNAP-25, t(16) = 2.652; p = 0.0174. (F) Quantifications of SNAP-25 volume not colocalized with syntaxin-3, t(16) = 0.508; p = 0.6182.

Unexpectedly, when we looked at syntaxin-3 and SNAP-25 colocalization in Munc18-1 deficient retinas, we noted a striking difference—synaptically targeted syntaxin-3 did not colocalize with SNAP-25 in the outer plexiform layer, rather, syntaxin-3 fluorescence appeared to be locating closer to the ONL when compared to the SNAP-25 signal (Fig. 3A). Using ImageJ, we generated intensity profiles for both syntaxin-3 and SNAP-25 in control and Munc18-1 cKO sections to quantitatively analyze the immunofluorescence signal (Fig. 3B). In control retina, the intensity profile of syntaxin-3 largely overlapped with the SNAP-25 signal and the peak of syntaxin-3 intensity was within the peak of SNAP-25 fluorescence. However, the syntaxin-3 peak shifts towards to the ONL relative to the peak of SNAP-25 in Munc18-1 deficient retinal regions (Fig. 3B). We additionally generated volume renderings of control and Munc18-1 cKO sections and confirmed a significant decrease in the volume overlap between syntaxin-3 and SNAP-25 (Fig. 3C,D). Furthermore, the proportion of syntaxin-3 located at the OPL that did not colocalize with SNAP-25 in Munc18-1 cKO was significantly increased in Munc18-1 deficient retinas (Fig. 3E), supporting the potential mislocalization of these two t-SNARE proteins. While we observed a slight narrowing of the SNAP-25 peak, the total volume of SNAP-25 signal alone was not significantly different between control or Munc18-1 cKO retina (Fig. 3F). Therefore, while the synaptic transport of syntaxin-3 remained largely intact in Munc18-1 deficient photoreceptor cells, the decreased colocalization between syntaxin-3 and its cognate t-SNARE partner suggests that in the absence of Munc18-1, there are alterations in t-SNARE protein localization.

The narrower syntaxin-3 band could also contribute to the decrease of postsynaptic syntaxin-3, as bipolar cells express syntaxin-3²⁴. Crx promoter drives expression in bipolar cells^22,29. Indeed, syntaxin-3 staining decreased in rod bipolar cells (Supplemental Fig. 6A,B). In the OPL, syntaxin-3 weakly co-stained with bipolar cell marker PKCα in the controls, but such staining was reduced in the Munc18-1 cKO (Supplemental Fig. 3B). Thus, a small reduction in postsynaptic syntaxin-3 likely contributed to the observed syntaxin-3 synaptic expression. However, as there was a strong decrease in syntaxin-3 and SNAP-25 colocalization in RIBEYE/PSD95-marked presynaptic areas (Supplemental Fig. 3A,B), a loss of postsynaptic syntaxin-3 cannot fully account for all the synaptic syntaxin-3 changes.

Munc18-1 is crucial for syntaxin-3 expression and localization, but not for other SNARE proteins in photoreceptors

Two possibilities exist for the loss of colocalization between syntaxin-3 and SNAP-25: first, there are changes in the presynaptic localization of SNAP-25. Loss of presynaptic SNAP-25 could potentially be accountable for the observed mismatch between syntaxin-3 and SNAP-25 localization. Second, despite syntaxin-3 being targeted to the synaptic region, syntaxin-3 does not sufficiently reaching the synaptic membrane where the photoreceptor active zone is. To rule out changes in SNAP-25, we first co-stained SNAP-25 with the ribbon marker RIBEYE. SNAP-25 was found in the OPL and colocalized with RIBEYE in control and Munc18-1 cKO photoreceptor terminals (Fig. 4A). Similarly, when SNAP-25 was co-stained with PSD95, marking the presynaptic terminal of photoreceptors, SNAP-25 continued to colocalize with PSD95 in the Munc18-1 cKO (Fig. 4B). Thus, we exclude the possibility that SNAP-25 presynaptic localization is altered. Next, we co-stained syntaxin-3 with RIBEYE in control sections (Fig. 4C). Syntaxin-3 was observed surrounding the signal of RIBEYE in the control synaptic terminals. However, in Munc18-1 cKO sections, while syntaxin-3 is present in photoreceptor synaptic terminals, the syntaxin-3 signal ends abruptly at the end of the region with RIBEYE fluorescence, and the syntaxin-3 signal did not extend beyond the RIBEYE signal (Fig. 4C).

Fig. 4.

Changes in SNARE localisation in Munc18-1 cKO is specific to syntaxin-3. (A) Confirmation that SNAP-25 presynaptic localization is unchanged in Munc18-1 cKO. SNAP-25 (green) remains strongly colocalized with presynaptic RIBEYE that marks the photoreceptor ribbon (red) in P14 Munc18-1 cKO retina. All images are maximum intensity projections. Scale bar = 20 μm. (B) Confirmation that SNAP-25 presynaptic localization is unchanged in Munc18-1 cKO. SNAP-25 (green) remains strongly colocalized with presynaptic PSD95 (red) in P14 Munc18-1 cKO retina. All images are maximum intensity projections. Scale bar = 20 μm. (C) Syntaxin-3 synaptic mislocalisation in P14 Munc18-1 cKO photoreceptors. In the control sections, syntaxin-3 (green) is distributed uniformly and diffusely throughout the whole of both rod and cone photoreceptors and syntaxin-3 signal is observed surrounding the RIBEYE (red) signal. However, in Munc18-1 cKO photoreceptors, syntaxin-3 is still in the synaptic terminals, but the accumulation is apical to the signal of RIBEYE and is more restricted. All images are maximum intensity projections. Scale bar = 20 μm. (D) Syntaxin-3 synaptic mislocalisation in P14 Munc18-1 cKO. Syntaxin-3 (green) staining pattern is altered in P14 Munc18-1 cKO retinas—while syntaxin-3 fills photoreceptor synaptic terminals identified by PSD95 (red) in both control and Munc18-1 cKO sections, PSD95 is within the milieu of the diffuse syntaxin-3 signal in control sections, while in Munc18-1 cKO sections, PSD95 and syntaxin-3 share a basal boundary that syntaxin-3 does not extend past. All images are maximum intensity projections. Scale bar = 20 μm.

When co-stained with the PSD95, we similarly observed syntaxin-3 in the presynaptic region in both control and Munc18-1 deficient retinas, yet again, the syntaxin-3 signal did not pass beyond the PSD95 signal (Fig. 4D). This loss of syntaxin-3 colocalization with SNAP-25 and syntaxin-3 being limited at RIBEYE was not an artifact of section angle as we observed this consistently for all mice tested (Supplemental Fig. 7). Therefore, while SNAP-25 localization appears unchanged in P12 Munc18-1 cKO retina, syntaxin-3 localization was closer to the ONL in the absence of Munc18-1. We also looked at the relative location of syntaxin-3 and the vesicular SNARE protein synaptobrevin/VAMP-2. In control sections, syntaxin-3 and synaptobrevin-2 colocalized in the presynaptic OPL, and we did not observe a dramatic shift in syntaxin-3 localization relative to synaptobrevin-2 in Munc18-1 cKO retinas (Supplemental Fig. 8). Syntaxin-3 and synaptobrevin-2 looked comparable in both control and Munc18-1 cKO sections. On the complementary, as mentioned previously, there could be a minor reduction of syntaxin-3 in postsynaptic cells (Supplementary Fig. 6), causing a decreased level of syntaxin-3 in the OPL. Overall, loss of Munc18-1 specifically affects the localization of syntaxin-3 at the synapse.

Munc18-1 knockout in photoreceptors leads to the loss of syntaxin-3 around photoreceptor nuclei

To further investigate syntaxin-3 mislocalization in the photoreceptors, we performed immunoelectron microscopy (immuno-EM) of syntaxin-3. The monoclonal syntaxin-3 antibody used above was a modified version used in the study of syntaxin-3 cKO in the retina⁹. However, this antibody did not work in immuno-EM. Thus, we utilized a different syntaxin-3 antibody (a rabbit polyclonal antibody), which was previously utilized to stain the inner segment membrane³⁰. We first verified that the new rabbit polyclonal syntaxin-3 antibody replicated the findings of decrease of syntaxin-3 in the outer nuclear layer of Munc18-1 cKO (Fig. 5A). This antibody also detected a colocalization of syntaxin-3 with ribeye in the Munc18-1 cKO (Fig. 5B). Puncta of syntaxin-3 were observed in cKO retina (Fig. 5C). However, unlike the monoclonal antibody, intensity of syntaxin-3 staining decreased in the OPL (Fig. 5D). Loss of colocalization between syntaxin-3 and SNAP-25 was replicated with the rabbit polyclonal antibody (Supplemental Fig. 9). The unaltered localization of syntaxin-3 with synaptic markers RIBEYE (Supplemental Fig. 10A), PSD95 (Supplemental Fig. 10B), and synaptobrevin-2 (Supplemental Fig. 10C) were all evident in Munc18-1 cKO, comparable to the previous findings. In addition, co-staining of syntaxin-3 with bipolar cells confirmed small postsynaptic syntaxin-3 reduction in the OPL (Supplementary Fig. 11A). No branching of horizontal or bipolar cells was observed in the ONL, confirming a lack of dendritic sprouting (Supplementary Fig. 11A,B). Thus, the reduced syntaxin-3 expression in photoreceptors was supported by two independent antibodies.

Fig. 5.

Verification of syntaxin-3 mislocalization by a different syntaxin-3 antibody. (A) Munc18-1 (green) and syntaxin-3 (rabbit polyclonal (pc) from ProteinTech) (red) stain in P12 control and Munc18-1 cKO retina. The loss of syntaxin-3 and punctate accumulation in the ONL (indicated by yellow arrowheads) were replicated with this antibody. (B) Syntaxin-3 (pc) (red) and RIBEYE (white) stain in P12 control and Munc18- 1 cKO retina. Similar finding on synaptic localization of syntaxin-3 using an independent antibody. All images are maximum intensity projections. (C) Punctate syntaxin-3 staining in the ONL. Two-tailed two-sample t-test. ONL: t(5) = − 5.31878; p = 0.00314. (D) Syntaxin-3 (pc) signal intensity is decreased in both ONL and the OPL (marked by the presence of RIBEYE). n = 3 in control, n = 4 in cKO. Two-tailed two-sample t-test. ONL: t(5) = 13.13238; p < 0.0001. OPL: t(5) = 3.6032; p = 0.01549.

Using the rabbit polyclonal syntaxin-3 antibody in immuno-EM, the control retina showed strong staining in the outer nuclear layer. Syntaxin-3 staining was observed in areas between photoreceptor nuclei (Fig. 6). Photoreceptor cell bodies were highly compact and the cell soma is small, where majority of cell body is composed of the nucleus. Previous EM pictures showed only small cytoplasmic spaces between nuclei in the OPL^31,32. Therefore, the staining pattern of syntaxin-3 in immuno-EM reflects cytoplasmic and plasma membrane staining in photoreceptors. We confirmed that syntaxin-3 had minimal co-staining with Müller glia (Supplementary Fig. 11C). Therefore, the signal was unlikely to be coming from Müller glial cells, but photoreceptors. In contrast, Munc18-1 cKO outer nuclear layer showed a drastic reduction of syntaxin-3 staining (Fig. 6). Therefore, loss of Munc18-1 in photoreceptors clearly impacted the proper expression of syntaxin-3 at the nuclear and cytoplasmic regions. The ultrastructural imaging further supports the importance of Munc18-1in determining the proper expression of syntaxin-3 in photoreceptors.

Fig. 6.

Immuno-electron micrograph revealing the decrease of Syntaxin-3 in photoreceptors. Immuno-electron micrograph staining of syntaxin-3 in P13 control and Munc18-1 cKO retina. Syntaxin-3 in the control (yellow arrowheads) is stained surrounding the photoreceptor nuclei (red) in the ONL. Syntaxin-3 expression is much lower in the Munc18-1 cKO (white arrowheads). Top panel: scale bar = 2 μm; middle and bottom panels: scale bar = 1 μm.

Outer segment and synaptic deficits in Munc18-1 cKO photoreceptors

We additionally did transmission electron microscopy to confirm changes at the ultrastructural level in P12 Munc18-1 control and cKO retinas. We observed skinny photoreceptor outer segments with neatly stacked membranous discs in the control sample, however, photoreceptor outer segments were often found to be swollen and vacuolated in the Munc18-1 cKO retina (Fig. 7A). These knockout outer segments also contained mitochondria and Golgi/endoplasmic reticulum appearing structures which are typically features of the photoreceptor inner segment (Fig. 7A). At the synaptic terminals, in the control, the synaptic boundary could be observed with developing photoreceptor synapses containing one synaptic ribbon per developing tripartite synapse. However, in Munc18-1 deficient photoreceptor cells, the synaptic boundary included vacuolated interruptions and dying photoreceptor nuclei could be observed (Fig. 7B). Furthermore, key synaptic features were altered—rather than a single synaptic ribbon per synapse, instances where multiple adjacent synaptic ribbons were present (Fig. 7C,D). Developing synapses in P12 control rod photoreceptor cells generally had 1 synaptic ribbon, although occasionally, 0 or 2 ribbons were observed. However, in Munc18-1 cKO rod photoreceptor cells, there were many instances of duplicated ribbons (Fig. 7D).

Fig. 7.

Outer segment and synaptic changes seen at ultrastructure level in Munc18-1 cKO retina. (A) Transmission electron micrographs of photoreceptor outer segments from P12 control and Munc18-1 cKO retina. Munc18-1 cKO outer segments are much more swollen and vacuolated compared to the skinny ciliary outer segments observed in controls. (B) Transmission electron micrographs of photoreceptor synaptic terminals. Developing photoreceptor terminals in control sections have many synaptic ribbons and clearly defined postsynaptic connections. Synaptic region (lined in red) has much more vacuolations and disturbances in Munc18-1 cKO retina. A loss of nuclear integrity is also observed in Munc18-1 cKO photoreceptors. (C) Control photoreceptor synapses have 1 synaptic ribbon per synapse while Munc18-1 cKO synaptic terminals often have multiple ribbons. Rod spherules are outlined in red. (D) Quantification of ribbons/rod spherule. A total of n = 237 rod spherules were quantified from N = 3 P12 control mice and n = 221 rod spherules from N = 3 P12 Munc18-1 cKO mice. A minimum of 50 rod spherules were scored for each animal examined. Munc18-1 cKO exhibited much more rod spherules with multiple ribbons. Two-sample two-sided t test; t₍₄₅₆₎ = 7.792, p < 0.0001, error bars represent SEM.

We also scored the synaptic ribbons of rod photoreceptor cells in control and Munc18-1 cKO retina. During development, the synaptic ribbon starts off spheroid, then elongates with a more “clubbed” shape, before reaching its final plate-like shape. The ribbon also becomes an anchor to the arciform density with development. P12 control photoreceptors generally had more mature ribbons that were anchored, with a small presence of immature ribbons (e.g., spherical ribbons, club-shaped ribbons, floating ribbons), however, Munc18-1 cKO retina had fewer anchored ribbons and more immature ribbons (Fig. 8A). We additionally characterized the state of rod synapses in control and Munc18-1 cKO rod photoreceptors. Photoreceptor synapses usually involve a synapse onto a pair of lateral horizontal cells and one to two bipolar dendrites at maturity. We quantified the postsynaptic elements present in rod spherules and found a dramatic decrease in the amount of mature triadic/tetradic connections in Munc18-1 cKO mice compared to their littermate controls (Fig. 8B). Many synapses lose one component, in which usually the horizontal connection is lost, while others lose all postsynaptic targets, yet remaining tethered to the arciform density (Fig. 8B). In TEM, we noted a larger vesicle size in the Munc18-1 cKO rod spherules. Together, we found the presence of synaptic deficits in Munc18-1 cKO mice, manifesting in the form of ribbon duplicates, increased presence of immature ribbons, and decreased mature photoreceptor synapses.

Fig. 8.

Munc18-1 cKO have more immature ribbons and less developed synapses. (A) Changes in synaptic ribbon state in P12 Munc18-1 cKO photoreceptors. Synaptic ribbons were quantified in P12 control and P12 Munc18-1 cKO mice according to their state: spherical, immature, and anchored. A total of n = 194 synaptic ribbons were quantified from N = 3 control mice and n = 333 synaptic ribbons from N = 3 P12 Munc18-1 cKO mice. A minimum of 50 rod spherules were scored for each animal examined. Ribbons were scored as either spherical, immature ribbons (free floating or clubbed), or anchored (with arciform density/opposed by at least two postsynaptic processes). Munc18-1 cKO exhibited a decrease in anchored ribbons and an increase in immature and spherical ribbons compared to control mice. Scale bar = 500 nm. (B) Changes in synapse state in P12 Munc18-1 cKO photoreceptors. A total of n = 47 rod spherules were scored from N = 3 P12 control mice and n = 55 rod spherules from N = 3 P12 Munc18-1 cKO mice. A minimum of 50 rod spherules were scored for each animal examined. Rod spherules were scored based on the presence and features of postsynaptic (PS) processes; a mature photoreceptor synapse (3–4 postsynaptic elements present), terminals with 1–2 postsynaptic elements, and empty terminals with 0 postsynaptic elements. Munc18-1 cKO rod spherules exhibited a decrease in mature synapses, and an increase in empty terminals and terminals with 1–2 postsynaptic elements when compared to controls. Scale bar = 500 nm.

Discussion

SNARE proteins play a crucial role in photoreceptor functions and survival, with evidence pointing towards the necessity of syntaxin-3 and SNAP-25 t-SNAREs in photoreceptors^9–11. However, the functional consequences of losing Munc18-1, the regulator of syntaxin and SNARE fusion, in the ribbon synapses and photoreceptors remain elusive. Here, we investigated the role of Munc18-1 in photoreceptor functions and survival, along with the Munc18-1 interactions with syntaxin-3 in photoreceptor cells. To achieve this, we utilized CRX-cre to remove Munc18-1 from photoreceptor cells at E12.5²², and then investigated changes in function, morphology, and protein expression. Removal of Munc18-1 led to decreases in retina thickness, largely between postnatal day 14–22 and diminished electroretinogram responses (Fig. 1). Notably, this 2–3-week window before Munc18-1 cKO retinas degenerated gave us the opportunity to investigate the interactions of Munc18-1 and syntaxin-3. In Munc18-1 deficient photoreceptors, we found changes to syntaxin-3 expression and localization – syntaxin-3 remained present in photoreceptor outer segments and was targeted towards synaptic terminals, but plasmalemmal and cytosol located syntaxin-3 in outer nuclear layer was significantly decreased (Fig. 2, supplemental Fig. 4). Decrease in overall syntaxin-3 expression in the ONL was coupled with aggregated puncta of synatxin-3 (Figs. 3 and 5). Despite syntaxin-3 partially being targeted towards synapses, we found a decrease in colocalization in synaptic syntaxin-3 relative to its cognate t-SNARE partner SNAP-25 at photoreceptor synapses (Figs. 3 and 5). Yet, the synaptic SNAP-25 and synaptobrevin-2 localization remained unchanged, and therefore, Munc18-1 removal selectively interfere with syntaxin-3 but not other SNAREs (Fig. 4). Using immuno-EM, we were able to confirm syntaxin-3 decrease around the photoreceptors cell soma (Fig. 6). Using transmission electron microscopy, we were also able to observe synaptic deficits before cell death and the early degeneration of photoreceptor outer segments (Figs. 7, 8 and 9).

Fig. 9.

Altered syntaxin-3 expression and photoreceptor structure in Munc18-1 deficient photoreceptors.

METHODS.

Our results indicate an essential role of Munc18-1 in photoreceptor cell survival and point toward potential regulation of outer segment dynamics and synaptic function. Recently, syntaxin-3 and SNAP-25 were determined to be the obligate t-SNARE proteins photoreceptor cells use in synaptic transmission and outer segment protein trafficking^9–11. Thus, we conclude that the photoreceptor ribbon synapses utilize not only SNARE protein but also sec/Munc18 protein dependent mechanisms for synaptic functions. The decrease in syntaxin-3 colocalization with SNAP-25 could potentially result in deficits in neurotransmitter exocytosis, leading to the impaired ERG responses observed at P14. Although the decline in ERG b-wave could be attributed to either a decrease in phototransduction or alterations in synaptic transmission, the extent of alterations to presynaptic neurotransmitter release warrants further investigation. However, due to the myriad of synaptic changes observed in Munc18-1 cKO mice (Figs. 7 and 8), it is certain that Munc18-1 does indeed contribute to synaptic organization and fidelity. We have previously generated SNAP-25 conditional knockout mice, using the same CRX-driven cre-recombinase that turns on at embryonic day E12.5; and SNAP-25 cKO photoreceptor cells have rapid degeneration starting at P9 that is complete by P14¹¹. This is in contrast to the Munc18-1 cKO mice, where photoreceptor cells are largely maintained until P14, but photoreceptors are lost by P21. Previously, syntaxin-3 conditional knockout mice have been generated using the same CRX-cre⁹. In syntaxin-3 cKO mice, photoreceptor degeneration is evident by P15, some photoreceptors remain at P21 although their outer segments are absent, and no photoreceptors remain by P30. Notably, despite having differences in time course, all three synaptic mutants exhibit a faster time course of degeneration when compared to other retinal mutants such as RD1 mice. Of course, we cannot exclude the contributions of syntaxin-3 in our Munc18-1 cKO mice, as indeed it has been proposed previously that in the absence of Munc18-1, syntaxin-3 confers negative consequences^33,34.

Similar to both previously reported SNAP-25 cKO and syntaxin-3 cKO, Munc18-1 cKO exhibited an increase in free-floating and immature ribbons, abnormal and accumulated vesicles, and abolished ERGs. These features may be characteristic of mutants underlying synaptic dysfunctions. We also observed aggregates of syntaxin-3 in the inner and outer segments of both Munc18-1 cKO and SNAP-25 cKO, suggesting that these ciliary occlusions may be an artefact of severe degeneration. Indeed, it has been proposed that the photoreceptor outer segment could act as a sink for membrane proteins³⁵. Membrane proteins would continue to localize to the outer segments even when protein trafficking mechanisms are impaired, and the continuous renewal of photoreceptor outer segments takes care of aberrant inclusions of erroneously targeted proteins to try to counter degeneration. These outer segment accumulations of SNAP-25 and Munc18-1 had also been previously been observed in syntaxin-3 knockout mice, where SNAP-25 and Munc18-1 became punctate accumulations that colocalized with outer segment proteins peripherin-2 and rhodopsin⁹. These common observations strengthened the necessity of vesicular-dependent trafficking and membrane fusion for the delivery of outer segment proteins. On the contrary, syntaxin-3 knockout in cones only depleted Munc18-1 expression in the inner segment and OPL³⁶. Future work looking into cone-specific removal of Munc18-1 could provide further insights into the interplay between Munc18-1 and syntaxin-3 in vesicular fusion of cones.

Syntaxin-3, particularly the syntaxin-3B isoform found in photoreceptors, is critical not just for vesicle fusion, but also for photoreceptor survival and proper trafficking of outer segment proteins^9,36,37. Since our Munc18-1 knockout disrupts syntaxin-3 expression levels, normal syntaxin-3 functions could be impaired. Similarities between our Munc18-1 cKO and the two previously reported syntaxin-3 knockouts could be due to common affected pathways. Further, syntaxin-3 dysfunction can underlie some of our observed phenotype, as our Munc18-1 cKO’s timeline for degeneration is much closer to that of syntaxin-3 photoreceptor-specific cKO rather than SNAP-25 cKO^9,11.

While syntaxin-3 and SNAP-25 are both enriched in control photoreceptor synaptic terminals, the colocalization of syntaxin-3 and SNAP-25 is reduced in the Munc18-1 deficient nerve terminals. This suggests that Munc18-1 is essential for the proper expression and localization of syntaxin-3 in the photoreceptor synapse (Fig. 8). Can this finding extend to more conventional neurons and synapses in the brain? In conventional synapses, rather than syntaxin-3, syntaxin-1 A and syntaxin-1B are the key isoforms for synaptic vesicle exocytosis. Our results showing that syntaxin-3 is partially transported to terminals of Munc18-1 deficient photoreceptors seem to be consistent with previous results that syntaxin-1 was correctly targeted to synapses in embryonic day 18 (E18) Munc18-1 null brain slices¹⁶. Using two different antibodies, the intensity of syntaxin-3 varies at the OPL, with the monoclonal antibody showing a wildtype level of staining and the polyclonal antibody showing a reduction (Fig. 5). The difference might be because the epitope targeted by the monoclonal antibody is more accessible when not binding or interacting with SNAP-25. Further, our finding of perinuclear aggregation of syntaxin-3 in the soma of photoreceptors is also consistent with cytoplasmic accumulation of syntaxin-1 in the soma of neurons that are differentiated from patient-derived Munc18-1 heterozygous nonsense mutant iPSCs¹⁹. Thus, we speculate that syntaxin-1 localization with respect to other SNAREs might be impacted in Munc18-1 deficient conventional synapses.

Although Munc18-1 has been recognized as a regulator of neurotransmitter exocytosis, its role is considered to extend beyond regulation of exocytosis. For instance, Munc18-1 removal results in the complete absence of both spontaneous and evoked release¹³, which is more severe than the phenotype of the knockout of synaptobrevin-2 or SNAP-25^38–41. While spontaneous release is still present in synaptobrevin or SNAP-25 knockouts, there is no spontaneous release in Munc18-1 knockout. The severe phenotype of Munc18-1 absence is more comparable to the double knockout of syntaxin-1 A and 1B, which is embryonic lethal^15,42. The essentiality of Munc18-1 homologues is also demonstrated in lower organisms such as Drosophila and C. elegans. Rop, the Drosophila homologue of Munc18, is essential for the life of Drosophila and its null mutation results in embryonic lethality⁴³. Meanwhile, the C. elegans homologue of Munc18 is known as UNC-18, and while unc-18 null C. elegans are viable, recent results suggest that this viability is supported by the presence of another isoform of Munc18, UNCP-18, and a double mutation of unc-18 and uncp-18 results in embryonic lethality⁴⁴. Such severe phenotypes observed in Munc18 mutants do not seem to be explained by the mere regulatory roles of Munc18-1 in the SNARE complex⁴⁵. Previous studies emphasise the regulatory role of Munc18-1 after t-SNARE complex assembly or in the acceleration of tertiary SNARE complex assembly^23,46–48, and such a regulatory role of Munc18-1 seem to be mediated by its domain 3^23,49–51. However, our results suggest that Munc18-1 is essential for syntaxin to be expressed and localized properly to execute its presynaptic functions in photoreceptors.

Our findings highlight the importance of the Munc18-1-syntaxin-3 interaction in the retina. Molecularly, Munc18-1 binds to syntaxin-3 to maintain it in the closed formation⁵². Recently, this interaction was spotlighted in a study in human patients with congenital nystagmus – the p.His16Arg mutation enhanced the binding interaction of Munc18-1 to syntaxin-3²⁰. Using two different antibodies, we found a reduction in the overall level of syntaxin-3 in the ONL, suggesting the necessity of Munc18-1 to maintain proper expression of syntaxin-3 in photoreceptors. Syntaxin is notorious for interacting with numerous proteins and is a very promiscuous protein^53–55; syntaxin-1 also has a known tendency to oligomerize and cluster, and this property is likely shared by syntaxin-3. Thus, without the protection of Munc18-1, it may aggregate with itself or bind with other synaptic proteins, hampering its arrival at the terminal plasma membrane. Though syntaxin-3 staining in the ribbon synapse will best address whether Munc18-1 is necessary for proper syntaxin-3 synaptic localization, we were unable to visualize syntaxin-3 staining with immuno-EM. There are yet studies showing immuno-EM of SNAREs in the synapse. The inability of detecting syntaxin-3 at the synapse could be because when syntaxin-3 form a SNARE complex, the accessibility of the epitope by the antibody was reduced after fixation used for immuno-EM. It has also been previously proposed that that syntaxin-3 has a lower affinity for SNAP-25^56,57, which may indicate syntaxin-3 may mostly exist in the closed conformation, though our results may suggest that the lowered affinity of syntaxin-3 for SNAP-25 may only be partially attributed to Munc18-1 stabilizing a closed conformation of syntaxin-3. Alternatively, Munc18-1 may provide the mechanisms to transfer syntaxin away from intramembrane structures to the terminal plasma membrane.

The aberrant syntaxin-3 synaptic localization was also caused partly by a reduction of syntaxin-3 expression in rod bipolar cells (Supplemental Fig. 6, 11A). Munc18-1 and syntaxin-3 are both expressed in bipolar cells⁵⁸. CRX-cre has the ability to drive knockout in photoreceptor and bipolar cells²². The decreased syntaxin-3 expression in postsynaptic regions could also contribute to the ERG functional deficits. More specific KO of syntaxin-3 and Munc18-1 in bipolar cells will explain how they work in the postsynaptic area or bipolar cell signal transmission.

Given the carefully controlled nature of vesicle release in photoreceptor cells of the retina, wherein the readily releasable pool and the reserve pool both need to be exquisitely primed, we highlight the necessity of identifying the precise mechanisms and components to prepare the SNARE complex for fusion. Beyond its location at the plasma membrane for vesicular fusion, syntaxin-3 has been identified in the vesicle fraction from protein pulldown experiments⁵⁶. Syntaxin-3 on synaptic vesicles has important implications for the multivesicular release present in photoreceptors. Munc18-1’s role in the preparation of vesicles for fusion and the localization of these proteins emphasize the importance of further investigation of synaptic proteins in the retina. Together, our findings suggest that Munc18-1 is necessary for the proper expression and localization of syntaxin-3 to the photoreceptors, and therefore, maintain functions and survival of photoreceptors.

Experimental model and subject details

Munc18-1 and SNAP-25 flox mice^11,23,59,60 of both sexes between postnatal day 9 and up to 2 months of age were used in this study. SNAP-25 flox and conditional mice were generated as previously described¹¹, briefly, For the flox mice, loxP sites were introduced to surround exon 5 of SNAP-25 in SNAP-25 flox mice and SNAP-25 flox mice were crossed with CRX-cre mice (https://www.informatics.jax.org/allele/MGI:5433291) to obtain SNAP-25 cKO mice^11,22. For Munc18-1 flox mice, loxP sites were introduced surrounding exon 7, and Munc18-1 conditional knockout mice were generated by crossing Munc18-1 flox mice with CRX-cre mice. Munc18-1 flox mice were obtained as ES cells from the European Conditional Mouse Mutagenesis (EUCOMM) Program, then the Munc18-1 flox mice were generated at the Toronto Center for Pharmacogenetics and maintained in house. Munc18-1 CRX-cre mice were kindly provided by Dr Valerie A. Wallace (University Health Network) and SNAP-25 mice were a kind gift from Dr. Weichun Lin (UT Southwestern). Genomic DNA was obtained from tail clippings using alkaline lysing methods and animals were genotyped using PCR (refer to Table 1 for primers). Tail clippings were lysed with 50 mM NaOH and incubated at 96 °C for 1 h with vigorous shaking. Littermate flox/flox mice without CRX-cre were used as controls for all experiments. All mice were maintained on a C57BL/6 background. We have previously tested our colony for the absence of contamination by Pde6b^{rd1 11}. All mice were housed in a vivarium with a 12-h light on/off cycle and temperatures were maintained between 22–23 °C. We have complied with all relevant ethical regulations for animal use. All animal experiment were in accordance with the ARRIVE Guidelines for the reporting of animal experiments. All experiments detailed here were reviewed and approved by the animal care committee of the University Health Network. All experiments were carried out in accordance with the Canadian Guidelines for Animal Care.

Table 1.

Reagents and resources.

Reagent or resource	Source	Identifier
Antibodies
Rabbit polyclonal Munc18-1	Proteintech	11459-1-AP
Mouse monoclonal Munc18-1	BD	CL 31
Mouse monoclonal SNAP-25	Covance	SMI-81R
Rabbit monoclonal Syntaxin-3	Abcam	133,750
Rabbit polyclonal Syntaxin-3	ProteinTech	15556-1-AP
Mouse monoclonal Calbindin	ThermoFisher	50-173-6822
Mouse polyclonal PKCɑ	Santa Cruz Biotechnology	sc-8393
Guinea Pig Glutamine synthetase	Synaptic Systems	367 005
Rabbit PSD95	Abcam	18,258
Mouse PSD95	Abcam	2723
Guinea Pig synaptobrevin-2	Synaptic Systems	104 204
Guinea Pig RIBEYE	Synaptic Systems	192 104
Mouse Synaptophysin	Sigma	S5768
Alexa Fluor 488 Goat anti-Rabbit IgG	Invitrogen	A-11008
Alexa Fluor 488 Goat anti-Mouse IgG	Invitrogen	A-11001
Alexa Fluor 568 Goat anti-Rabbit IgG	Invitrogen
Alexa Fluor 568 Goat anti-Mouse IgG	Invitrogen
Alexa Fluor 647 Goat anti-Rabbit IgG	Abcam
Alexa Fluor 647 Goat anti-Mouse IgG	Abcam
Alexa Fluor 647 Goat anti-Guinea Pig IgG	Invitrogen
Goat anti-Mouse Secondary Antibody, IgG HRP-conjugated	Jackson ImmunoResearch	115-035-003
Nanogold-Fab” goat anti rabbit igG (H + L)	Nanoprobes, Inc.	2004
HQ Silver TM	Nanoprobes, Inc	2012
ECL™ Western Blotting Reagents	Cytiva	GERPN2106
Ponceau S	Sigma	P3504
Chemicals, peptides, and recombinant proteins
Taq DNA polymerase	Bio-Helix	Cat #. MB101-0500
Euthanyl	Bimeda-MTC	DIN 00141704
Paraformaldehyde	EM Science	CAS 30525-89-4
PBS Tablets	BioShop	Cat #. PBS404.200
Tissue-Tek O.C.T. Compound	Sakura	Cat #. 62550-12
Triton X-100	Sigma	CAS 9002-93-1
Goat Serum	Gibco	Cat #. 16210-072
Mowiol 4–88	Sigma	Cat #. 81381-250
In Situ Cell Death Detection Kit, Fluorescein	Roche	11,684,795,910
Experimental models: Organisms/strains
Mouse: Stxbp1tm1c(EUCOMM)Hmgu (Munc18-1 flox)	EuCOMM23	N/A
Mouse: B6-Snap25tm3mcw (SNAP-25 flox)	11,59
Mouse: Tg(Crx-cre)764Gla (CRX-cre)	22
Mouse: Stxbp1tm1c(EUCOMM)Hmgu; Tg(Crx-cre)764Gla (Munc18-1 cKO)	This paper	N/A
Mouse: B6-Snap25tm3mcw; Tg(Crx-cre)764Gla (SNAP-25 cKO)	11	N/A
Oligonucleotides
5’CGA ACT TCC ACA GCC CTT TAC AGA 3’	This paper	(1287) Munc18-1 (Forward)
5’GTG CAT GTG TAT GCC TGG ACT CAC 3’	This paper	(1288) Munc18-1 (Reverse)
5’- CCCTGGGGAACCACGGCAGA − 3’	11	(1911) SNAP-25 (Forward)
5’- TCCCAGGAAACAGCACAGCGT − 3’	11	(1912) SNAP-25 (Reverse)
5’ - AAACCTGAAGCAGGATGTGAGT − 3’	11	(1919) Cre (Forward)
5′- ACAGAAGCATTTTCCAGGTATGCT-3′	11	(1747) Cre (Reverse)
Software and algorithms
Adobe photoshop	Adobe	www.adobe.com
Adobe illustrator	Adobe	www.adobe.com
ImageJ	National institute of health	Imagej.net/software/fiji
Espion	Diagnosys	Diagnosysllc.com
Spectralis	Heidelberg Spectralis	Heidelbergengineering.com
OptoMotry	CerebralMechanics	www.cerebralmechanics.com
Zen Blue	Carl Zeiss	Zeiss.com
Imaris 10.1
Prism 9.5.1	Graphpad
Other
Microscope Cover Glass	Fisherbrand	Cat #. 12545 M
Superfrost Plus Slides	Fisherbrand	Cat #. 12-550-15

Preparation of retinal samples for immunohistochemistry

Mice were anaesthetized with an intraperitoneal injection of sodium pentobarbital (75 mg/kg, Bimeda-MTC) and subsequently euthanized. Eyes were placed in 4% PFA following enucleation. Using a 32G syringe, a hole was poked in the cornea and the front of the eye was removed to assist with the fixation process, the retina and all extra scleral tissue was left intact. Eyes were postfixed for 2 h at room temperature in 4% PFA. Following, three PBS washes, the eyes were placed in 30% sucrose overnight at 4 °C for dehydration. The next day, eyes were mounted in OCT and stored at − 80 °C until sectioning. 14 μm sections were obtained using a Leica CM3050S cryostat. Sections were dried and stored at − 80 °C until use.

Preparation of retina samples for transmission electron microscopy

Mice were euthanized as previously described using sodium pentobarbital (75 mg/kg) and transcardially perfused with 5–10 mls of PBS followed by EM fixative. Eyes were enucleated and placed into EM fixative. The front of the eye was removed but the retina and all extra scleral tissue was left intact. An approximately 4 mm × 4 mm square was cut from the mid-periphery of the eye for further processing. Further sample processing for transmission electron microscopy were performed by the Nanoscale Biomedical Imaging Facility, The Hospital for Sick Children, Toronto, Canada. Transmission electron micrographs were taken on a Hitachi HT7800 transmission electron microscope.

Preparation of retina samples for pre-embedding immuno-electron microscopy

Mice were euthanized using sodium pentobarbital (75 mg/kg) and transcardially perfused with 5–10 mls of PBS followed by 5–10 mls of fixative solution (4% PFA + 0.1% glutaraldehyde) for 3–5 min. Eyes were enucleated and placed into fixative. A 32G syringe was used to poke a hole in the cornea and an opening was cut in the front of the eye to allow fixative to enter the eye cup. The lens was removed and the eyecup was fixed for 1 h. The eyecup was then washed three times using PBS for 5 min each. The eyecup was embedded in low-gelling temperature agarose and 150 μm sections was cut using a vibratome and PBS as a buffer. Retina sections were collected into a well plate with PBS until further use.

Sections were permeabilized and blocked using 10% goat serum + 0.05 triton X-100 in PBS on a shaker for 30 min. After 30 min, the blocking solution was removed and the primary antibody solution (1:1000 syntaxin-3 antibody in PBS + 1% goat serum + 0.005% triton X-100 was added to the sections. Sections were incubated in primary antibody solution with gentle shaking at 4 °C for 4 days.

After primary incubation, sections were washed 3 times in PBS + 1% goat serum for 5 min each. Secondary antibody solution (1:100 anti-rabbit 1.4 nm nanogold in PBS + 1% goat serum + 0.005% triton X-100 or 1:100 anti-rabbit 6 nm nanogold in PBS + 1% goat serum + 0.005% triton X-100) was applied and sections were incubated for 2 h. Sections were then washed twice with PBS + 1% goat serum, and twice more with PBS. Finally, sections were placed in fixative and stored until further processing. For 1.4 nm gold samples that were silver enhanced, samples were postfixed for 30 min, then washed 3 times in PBS for 5 min each. Another 3 washes in water for 5 min followed. Silver enhancement was performed according to instructions provided by Nanoprobes, briefly, initiator, moderator, and activator reagents were mixed thoroughly. Then, the solution was applied to the sections and the tissue was developed in the silver enhancement mixture for 2.5 min. Afterwards, the tissue was washed twice with cold water washes for 5 min each wash. The enhanced sections were placed in fixative and stored until further processing. Further sample processing for transmission electron microscopy were performed by the Nanoscale Biomedical Imaging Facility, The Hospital for Sick Children, Toronto, Canada. Transmission electron micrographs were taken on a Hitachi HT7800 transmission electron microscope.

Immunohistochemistry

Retina sections were washed three times for 3 min each in PBS followed by three 5-minute washes in PBS with 0.1% triton X-100 (PBS-Tx). Sections were then blocked with 10% goat serum in PBS-Tx for 1 h. Primary antibody was diluted 1:500 in blocking solution and left overnight at 4 °C in a humidified chamber. The following day, five by 5-min PBS-Tx washes were done, and the secondary antibody solution (1:1000 antibody in with DAPI + PBS-Tx) was applied to sections for 1 h at room temperature and covered from light. Further antibody detail and information are provided in the attached resource table (Table 1). Sections were then washed five times using 5-min PBS-Tx washes, while covered from light and with shaking before sections were placed in PBS and mounted in Mowiol.

Immunofluorescence imaging

Sections were imaging using a Carl Zeiss confocal microscope (LSM780) and with a minimum averaging of 4×. Images were taken at 20× and 63× using 1.0 Airy Unit pinhole sizes. For 63× oil immersion images, Zeiss immersion oil 518 F was used. A region 0.5–1.0 mm away from the optic nerve was selected for imaging and displayed in all representative immunofluorescence images.

Image processing and quantification

For profile analysis, line intensity histograms were generated in ImageJ. For each image, three regions were selected and within each region, three 5 μm wide line profiles were generated. The intensity of the syntaxin-3 and SNAP-25 channels were generated along each pixel of the line. The generated profiles were then aligned to the peak intensity of syntaxin-3. The three 5 μm wide line profiles within each region was then averaged and the resulting separated signal intensities for each dye were plotted as a function of distance.

For volume analysis, images were rendered in Imaris 10 (Bitplane, Zurich). Using the surfaces function, a surface of a 500 px wide, 100 px tall region of the outer plexiform layer was generated for both syntaxin-3 and SNAP-25 channels.

Optical coherence tomography

Spectral-domain optical coherence tomography scans were performed on mice of 3 weeks of age using the Heidelberg Spectralis animal imaging system. Mice were anesthetized using Avertin (125–250 mg/kg) and had their eyes dilated using Mydriacyl. The left eye of each mouse was used for all scans and all thickness measurements were taken 1.0–1.5 mm away from the optic nerve head, automated segmentation provided by the Heidelberg Spectralis animal imaging software and the experimenter manually fixed erroneous segmentation. Each round of scanning was performed in under 10 min and mice were let to recover on a heated pad.

Electroretinograms

Electroretinogram recordings were performed on P22 mice. Light stimulation was through the Diagnosys Espion electroretinogram device with a ColorDome Ganzfeld stimulator and Espion software. Before recordings, mice were dark adapted overnight and all recordings were done under red illumination. Mice were anesthetized using Avertin (125–250 mg/kg). Mydriacyl (Alcon) eye drops were used to dilate the eyes, and topical Alcaine (Alcon) was also applied. Body temperature was maintained at 38 °C throughout examination using the Diagnosis device. Recordings were taken from both eyes using 2 mm gold loop electrodes placed on the cornea of the eyes. Reference electrode was placed between the eyes subcutaneously and a ground electrode in the tail. The cornea was kept moist with a thin layer of methylcellulose solution. Light stimulation consisted of brief pulses of white light (6500 K). Scotopic recordings were done first on the dark-adapted mouse. 10 recordings per stimulus intensity were averaged, and the a- and b wave amplitudes were manually determined by the experimenter. Scotopic stimuli conditions are as follow: 0.00025 cd/s/m², 0.0025 cd/s/m², 0.025 cd/s/m², 0.25 cd/s/m², 2.5 cd/s/m², 5 cd/s/m², and 10 cd/s/m². Mice were then light adapted with 30 cd/s/m2 background illumination for 10 min. Photopic tests were done in the presence of 30 cd/s/m² background illumination. 10 recordings per flash intensity were averaged, and the amplitude of the a- and b waves was manually determined by the experimenter again. The following photopic stimulus intensities were used: 0.15 cd/s/m², 1.25 cd/s/m², 5 cd/s/m², 10 cd/s/m², 25 cd/s/m², 50 cd/s/m², and 75 cd/s/m².

Optomotry testing

Optomotry testing was performed on Munc18-1 control and cKO mice of 2 months of age. Quantification of visual acuity was performed using the OptoMotry system from CerebralMechanics. Vertical sine wave gratings were presented by a chamber composed of four monitors to create a virtual cylinder in the visual field of the mouse. Visual acuity is reported as the maximum spatial frequency (cycles/degree) at which head movements can still reliably be elicited. A Zeiss video camera was used to image the mouse from above. The mouse’s head was manually tracked by a software and the vertical sine wave gratings were then rotated around the mouse at 12 degrees/second using the randomized protocol design and the simple staircase psychophysical method within the software. The grating contrast was maintained at 100% resulting in a black/white pattern. Testing was done under normal lighting conditions and the luminance inside the OptoMotry chamber measured ~ 200 lux⁶², in the photopic range. The tester was blinded to spatial frequencies being tested during the procedure.

Western blot

Mice were euthanized with sodium pentobarbital (75 mg/kg) and transcardially perfused with 10mL PBS. Eyes were enucleated and lens were removed. For retina sample, the retina was collected by removing it from the eye cup. The remaining parts were considered as retinal pigment epithelium (RPE). Whole eye sample contained both the retina and the eye cup. Samples were homogenized by a sonicator. Protein concentration was measured with Bradford assay. 10 µg of samples were loaded into each lane in a 12.5% SDS-PAGE. Transfer was done at 4 °C onto a 0.45 μm nitrocellulose membrane for 2 h. Total protein level was inspected by 0.5% Ponceau S staining. Membrane was blocked with 5% skim milk dissolved in Tris-buffered saline with 0.1% Tween (TBS-T). Munc18-1 antibody was applied at a dilution of 1:5000 at 4 °C overnight. After TBS-T washes, membrane was probed with anti-mouse horseradish peroxidase (HRP)—conjugated antibody for 1 h. After TBS-T washes, membrane was probed with ECL™ Western Blotting Reagents and imaged in Bio-Rad Gel Doc XR + Gel Documentation System.

Statistics and reproducibility

All statistical analyses were done in Prism 9.5.1. Independent t-tests were used for two-group experiments with a p-value < 0.05 as the threshold for statistical significance. For all quantified tests (OCT, ERG, OKT), minimum triplicates were used at every time point. No data were excluded from analysis. Reproducibility was confirmed using triplicates. Two-way analysis of variance (ANOVA) was used for comparisons of multiple groups, with a significance level of 0.05. All error bars represent SEM.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

M.H.: methodology, investigation, writing—original draft, review, and editing, visualization. C.H.C.: methodology, investigation, writing—original draft, review and editing, visualization. R.N.: methodology, investigation, writing-review and editing, visualization. J.C., K.I., H.H.: methodology, investigation, writing—review and editing. V.A.W.: conceptualization, resources, writing—review and editing. P.P.M.: resources, writing—review and editing. S.S.: conceptualization, writing—original draft, supervision.

Data availability

This study did not generate new code. The datasets analysed during the current study are available from the corresponding author upon request.

Materials availability

All reagents and mouse lines generated by this study are available from the lead contact.

Declarations

Competing interests

The authors declare no competing interests.