(A) Expression patterns of strip1 mRNA during embryonic development. In situ hybridization using a strip1 antisense probe is shown at different developmental stages: 4 cell, blastula, 75% epiboly, 1, 2, and 3 dpf. In situ hybridization using the sense probe at 1 dpf is shown as a negative control. strip1 mRNA is ubiquitously expressed until gastrula stage. At 1–2 dpf, expression becomes restricted in the heart, optic tectum, and eyes. Ventral view of 2-dpf retina shows a stronger signal in GCL (left panels). OT, optic tectum; GCL, ganglion cell layer. Scale bar, 100 μm.(B) Frontal cryosections of 2-dpf heads following in situ hybridization show that strip1 mRNA is predominantly expressed in the inner retina and the optic tectum. (C) Retinal cryosections show higher expression in RGCs and ACs. Right panels show higher magnification of outlined areas. RGCs, retinal ganglion cells; ACs, amacrine cells; HCs, horizontal cells; PRs, photoreceptors. (D) Live confocal snapshots of the same wild-type and strip1^rw147 mutant retinas at different developmental stages; 40, 52, 62, and 76 hpf stained with Bodipy TR. A primitive IPL is first observed in wild-type retinas as early as 52 hpf. However, it is not as prominent in strip1^rw147 mutant retinas. By 62 hpf, strip1^rw147 mutants display severe defects in IPL formation. Arrowheads indicate rudimentary IPL. Asterisks show IPL defects in strip1^rw147 mutants. Scale bar, 50 μm.

(A) Differential interference contrast images showing lateral view of wild-type and strip1^rw147 mutant retinas. White dashed lines depict lens outline. Green dashed lines depict eye outline. Retina area is the area calculated between them. Scale bar, 50 μm. (B) Quantification of total retina area in wild-type and strip1^rw147 mutants. Retina area is slightly but significantly reduced in mutants, compared to wild-type siblings. Mann–Whitney U-test, n ≥ 4. (C) Live confocal images of wild-type and strip1^rw147 mutant retinas at 3 dpf combined with the transgenic line Tg[ptf1a:mCherry-CAAX], which labels ACs. Middle and lower panels show higher magnification of outlined areas. Abnormal positioning of ACs in the ganglion cell layer (GCL) is more prominent in areas of strong IPL defects than in areas of weak lamination defects. Scale bar, 50 μm. (D) Live confocal images of retinas of STD-MO and MO-ath5-injected wild-type embryos combined with the transgenic line Tg[ath5:GFP; ptf1a:mCherry-CAAX], which labels RGCs and ACs. Middle and bottom panels show higher magnification images to show abnormal localization of ACs in GCL (arrowheads) and disrupted IPL. Scale bars, 50 μm (upper panels) and 20 μm (middle and lower panels). (E) Wild-type and strip1^rw147 mutant retinas at 3.5 dpf labeled with anti-parvalbumin which labels subsets of ACs. Arrows indicate abnormal positioning of ACs in the GCL. Nuclei are stained with Hoechst. Scale bar, 50 μm. (F) The number of parvalbumin+ cells per retina. Student’s t test with Welch’s correction, n ≥ 3. (G) Percentage of parvalbumin+ cells (GCL+ or INL+) to the total number of parvalbumin+ cells. Two-way analysis of variance (ANOVA) with the Tukey multiple comparison test, n ≥ 3. For all graphs, data are represented as means ± standard deviation (SD). ns, not significant, *p < 0.05, **p < 0.01, and ****p < 0.0001.

Figure 2—figure supplement 1—source data 1.

Labeling of wild-type and strip1^rw147 mutant retinas at 4 dpf with zpr3 (A), zpr1 (B), anti-glutamine synthetase (GS) (C), and at 3 dpf with anti-PCNA (D) antibodies, which visualize rod photoreceptors, double-cone photoreceptors, Müller glia, and proliferative cells in the CMZ, respectively. Both rods and cones are in outer nuclear layer (ONL) and are mostly intact in strip1^rw147 mutants. Müller glia genesis and retinal stem cells and progenitor cells in the CMZ appear to be unaffected in strip1^rw147 mutants. Scale bar, 50 μm.

(A) The number of transferase dUTP nick end labeling (TUNEL+) in the ganglion cell layer (GCL) compared to the rest of the retina at 36, 48, 60, 72, and 96 hpf in strip1^rw147 mutant retinas, for experiments shown in Figure 3A. Apoptosis gradually increases to reach its peak at 72 hpf in other retinal areas, but its number at 72 hpf is significantly lower than that of the GCL at 72 hpf. In addition, there is no significant difference in apoptotic cell number of other retinal areas between 72 hpf and other time points (36, 60, and 96 hpf). Two-way analysis of variance (ANOVA) with the Tukey multiple comparison test, n ≥ 3. (B) Percentage of GCL area (area between the lens and inner plexiform layer [IPL]) relative to the total retinal area in wild-type and strip1^rw147 mutant retinas at 36, 48, 60, 72, and 96 hpf, for experiments shown in Figure 3A. There is no significant difference between wild-type siblings and strip1^rw147 mutants at any stage. Two-way ANOVA with the Tukey multiple comparison test, n ≥ 3. (C) Live confocal images of wild-type and strip1^crisprΔ10 mutant retinas stained with acridine orange (AO) to label apoptotic cells at 60 hpf. Scale bar, 50 μm. (D) Percentage of AO area relative to total retina area in 60 hpf wild-type and strip1^crisprΔ10 mutant retinas. Mann–Whitney U-test, n = 5. (E) TUNEL of 60 hpf wild-type and strip1^rw147 mutant heads combined with Tg[ath5:GFP] to label RGCs. All nuclei are counterstained with Hoechst. Apoptosis occurs markedly in RGCs and optic tectum in strip1^rw147 mutants. Scale bar, 100 μm. (F) The number of TUNEL+ in the optic tectum of wild-type and strip1^rw147 mutants at 60 hpf. The number of TUNEL+ cells is significantly higher in strip1^rw147 mutants than in wild-type siblings. Student’s t-test with Welch’s correction, n = 3. (G) Dorsal view of the optic tectum of 3-dpf wild-type and strip1^rw147 mutants. RGC axons are labeled using intraretinal injections of DiO or DiI. Axons of strip1^rw147 appear to exit the optic disc normally and the optic chiasm is formed. However, the optic nerve is thinner in strip1^rw147 mutants than in wild-type siblings. In addition, most axons fail to elongate properly and do not arborize within the optic tectum (arrowheads) or very few axons elongate poorly to reach more posterior arborization fields within the optic tectum (asterisk). Scale bar, 50 μm. For all graphs, data are represented as means ± standard deviation (SD). ns, not significant, *p < 0.05, and ***p < 0.001.

Figure 3—figure supplement 1—source data 1.

(A) Schematic drawings of cell transplantation design to evaluate cell autonomy of Strip1 in RGC dendritic patterning. Donor cells from strip1^rw147 heterozygous intercross carrying the transgene ath5:GFP (green) are labeled with dextran rhodamine (magenta) and transplanted into wild-type host embryos. (B) Maximum projection images of wild-type donor cells transplanted into wild-type host retinas as outlined in (A), arrowheads depict RGC axons. Scale bar, 10 μm. (C) Maximum projection images of strip1^rw147 mutant donor cells transplanted into wild-type host retina. Asterisks show dendritic projection defects, red arrowheads depict RGC axon defects. Scale bar, 10 μm.

(A) 3D confocal live images showing neurite projection of ACs to the IPL in transplantation assays outlined in Figure 4B, C. Panels represent three combinations of transplantation experiments: from wild-type donor to wild-type host (left columns), from strip1^rw147 mutant donor to wild-type host (middle columns), and from wild-type donor to strip1^rw147 mutant host (right columns). Donor cells are labeled with dextran-Alexa-488 in green and ACs are labeled with ptf1a:mCherry-CAAX in magenta. Mutant donor ACs show a normal dendritic pattern and normal projection to the IPL in wild-type host retina, although on rare occasion we observed that mutant ACs display two dendritic trees (arrow in the bottom middle panel), but still contribute to a normal IPL. On the other hand, wild-type donor ACs transplanted into mutant host retinas show abnormal projection patterns and infiltrate the ganglion cell layer (GCL; arrowheads in the right panels). Scale bar, 20 μm. (B) Confocal image of 3-dpf live wild-type and strip1^rw147 mutant retinas combined with the transgenic line, xfz43, to label subsets of BPs. Scale bar, 10 μm. (C) Schematic drawings of cell transplantation design to evaluate cell autonomy of Strip1’s role in BP development. strip1^rw147 mutant donor cells carrying the transgene xfz43 (green) are labeled with dextran-cascade blue or dextran-Alexa-flour 647 (blue) and transplanted into wild-type host embryos at blastula stage, leading to chimeric wild-type retina with strip1^rw147 mutant retinal columns. Retinal lamination of the host retina is visualized with Bodipy TR (magenta). (D) 3D confocal live images of four wild-type host retinas with strip1^rw147 mutant donor retinal columns outlined in (C). strip1^rw147 mutant donor BPs show normal neurite projection patterns to the IPL in wild-type host retina. Occasionally, bipolar axon terminals are abnormally extended laterally (arrow in the bottom panel); however, such defects are confined to a seemingly normal IPL. Scale bar, 10 μm. (E) Confocal image of 3-dpf live wild-type and strip1^rw147 mutant retinas combined with the transgenic line, xfz3, to label subsets of BPs. Scale bar, 10 μm. (F) Schematic drawings of cell transplantation design to evaluate the cell autonomy of Strip1’s role in BP development. Wild-type donor cells carrying the transgene xfz3 (green) are labeled with dextran rhodamine (magenta) and transplanted into strip1^rw14 mutant host embryos at blastula stage, leading to chimeric strip1^rw147 mutant retinas with wild-type retinal columns. (G) 3D confocal live images of two different strip1^rw147 mutant host retinas with wild-type donor retinal columns outlined in (F). Wild-type donor BPs show misguided neurite projections. Scale bar, 10 μm.

(A) Schematic diagram showing the core components of the mammalian STRIPAK complex and some of its accessory molecules. Core STRIPAK components include serine/threonine-protein phosphatase catalytic subunit known as PP2Ac and the scaffolding subunit PP2Aa, together with striatins (STRN, STRN3, and STRN4), Mob3, STRIP1, or STRIP2, and a germinal center kinase, GCKIII kinase, bound via cerebral cavernous malformation 3 (Ccm3). Some molecules bind in a mutually exclusive pattern to the core components. Arrows show that sarcolemmal membrane-associated protein (SLMAP) and the suppressor of IKKe (SIKE) are not detected in STRIPAK complexes containing cortactin-binding proteins (CTTNBP2/NL) and vice versa. (A has been adapted from Figure 3 in Hwang and Pallas, 2014). (B) STRING protein–protein interaction (PPI) network analysis of six zebrafish Strip1-interacting partners significantly enriched from IP-MS analysis highlighted in Figure 5C. Network edges represent known and/or predicted functional interactions in the STRING database. Edge thickness reflects the combined STRING evidence score for each binary relationship. Thicker edges represent increased interaction evidence. PPI enrichment p value <1.0e−16. (C) Uniform Manifold Approximation and Projection (UMAP) plots showing different clusters of retinal cells at 48 hpf analyzed with Seuret R pipeline using data published by Xu B. et al. (2020). Twelve clusters are identified and categorized to different retinal cell types or subtypes. RPCs, retinal progenitor cells; RGCs, retinal ganglion cells; ACs, amacrine cells; BPs, bipolar cells; HCs, horizontal cells; PRs, photoreceptors; MGs, Müller glia. UMAP plots showing retinal mRNA expression patterns of strip1 (D) and its interacting STRIPAK components, identified from Co-IP/MS; cttnbp2 (E), ttnbp2nlb (F), strn (G), strn3 (H), and strn4 (I). Only strip1 and strn3 show high retinal expression levels at 48 hpf compared to other partners.

List of enriched GO terms in DEGs downregulated (A) and upregulated (B) in strip1^rw147 mutant relative to wild-type siblings as analyzed with Metascape, FDR < 0.05 and log₂ FC > |1|. Downregulated genes show many GO terms related to the nervous system, RGCs, embryo morphogenesis, synaptic development, and transmission. Upregulated genes show terms related to stress response, cell death and MAPK signaling. (C) Venn diagram showing overlap of upregulated DEGs in RNA sequencing datasets of strip1^rw147 mutant relative to wild-type (FDR < 0.05 and FC > 1.5) in comparison with published upregulated DEGs in microarray data of adult zebrafish RGCs at 3 days post-optic nerve injury (ONI) (Veldman et al., 2007) and adult zebrafish eyes at 4 days post-optic nerve crush (ONX) (McCurley and Callard, 2010).

(A) Whole-mount labeling of 48-hpf wild-type and strip1^rw147 mutant retinas with anti-phospho-Jun (p-Jun) antibody and zn5 antibody, which label active Jun and RGCs, respectively. Arrowheads show RGCs expressing phospho-Jun in the ventronasal region. Scale bar, 50 μm. (B) Percentage of phospho-Jun area relative to zn5 area at 48 hpf. Mann–Whitney U-test, n ≥ 4. Data are represented as means ± standard deviation (SD). **p < 0.01.

Figure 6—figure supplement 2—source data 1.

(A) 60-hpf wild-type and strip1^rw147 mutants combined with the transgenic line Tg[hsp:mCherry-Bcl2]. Nontransgenic embryos (Bcl2−, top panels) are compared to transgenic embryos (Bcl2+, bottom panels) after heat shock treatment. Apoptotic cells are visualized by transferase dUTP nick end labeling (TUNEL) FL and fluorescent signals from mCherry-Bcl2 are shown. Nuclei are stained with Hoechst. (B) The number of TUNEL+ cells in ganglion cell layer (GCL). Two-way analysis of variance (ANOVA) with the Tukey multiple comparison test, n = 3. (C) 78-hpf wild-type and strip1^rw147 mutant retinas combined with the transgenic lines, Tg[ath5:GFP] and Tg[hsp:mCherry-Bcl2]. Nontransgenic embryos (Bcl2−, top panels) are compared to transgenic embryos (Bcl2+, bottom panels) after heat shock treatment. RGCs are labeled with ath5:GFP and fluorescent signals from mCherry-Bcl2 are shown. Anti-acetylated α-tubulin labels the IPL. Nuclei are stained with Hoechst. Arrowheads represent areas where RGC dendrites contribute to the IPL. Asterisks denote areas where RGC dendrites fail to project to the forming IPL and a fraction of presumptive amacrine cells is located between them. (D) Percentage of ath5+ area relative to retinal area. Two-way ANOVA with the Tukey multiple comparison test, n ≥ 3. Scale bar, 50 μm (A, C). For all graphs, data are represented as means ± standard deviation (SD). ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Figure 7—source data 1.

(A) In wild-type retina, Strip1 suppresses Jun-mediated proapoptotic signals, probably through the STRIPAK complex, to maintain retinal ganglion cells (RGCs) during development. In the absence of Strip1, Jun is activated in RGCs leading to severe degeneration of RGCs as early as 2 dpf. Subsequently, cells in the inner nuclear layer (INL) abnormally infiltrate the ganglion cell layer (GCL) leading to a disrupted inner plexiform layer (IPL). (B) Proposed model for Strip1’s role within RGCs to regulate amacrine cell (AC) positioning and IPL formation. In wild type, Strip1 regulates (1) RGC survival to prevent AC infiltration, and (2) RGC dendritic patterning to promote RGC–AC interactions. In strip1^rw147 mutants, both mechanisms are perturbed, leading to AC infiltration, increased AC–AC interactions, and IPL defects. In Bcl2-rescued strip1^rw147 mutants, survived RGCs prevent AC infiltration. However, RGC dendritic defects lead to increased AC–AC interactions and ectopic IPL formation.