Fig 1: Generation and confirmation of zebrafish Cralbp CRISPR knockout lines. (A) Schematic of CRISPR approach used to generate rlbp1a and rlbp1b knockout lines. (B) Representative genotyping gels for PCR genotyping of rlbp1a and rlbp1b knockout lines where amplicons of 346 bp and 245 bp (rlbp1a) or 469 bp and 248 bp (rlbp1b) represent wildtype and deletion alleles, respectively. (C) Brightfield images of rlbp1a, or rlbp1b, and double knockout lines. (D) Cralbp immunostaining in 5 dpf larval zebrafish eyes showing expression in Müller glia and RPE. (E) Cralbp immunostaining in the rlbp1a knockout line showing loss of expression in Müller glia and robust expression in the RPE. (F) Cralbp immunostaining shows the opposite pattern in the rlbp1b knockout line with loss of expression in the RPE but persistent expression in the Müller glia. Magenta staining represents staining with the RLBP1 antibody (15356‐1‐AP), blue represents staining with Hoechst for nuclei.
Fig 2: The EGFP‐tagged zebrafish rlbp1b gene but not the human RLBP1 gene can rescue the dim light OKR in rlbp1b knockout larvae. (A) Schematic of standard raising and dim light OKR approach used to assess rescue in humanized zebrafish lines. (B) Schematic of expression vectors used to create humanized EGFP‐tagged RLBP1 zebrafish lines used in this study. (C) The EGFP‐tagged human RLBP1 gene is not sufficient to rescue the dim light OKR phenotype in rlbp1b knockout larvae (p < 0.0001, Mann Whitney test). Three biological replicates n ≥ 16 larvae per group. (D, E) The expression of missense RLBP1 disease variants p.R151Q and p.A237V do not cause any additional impairment of the dim light OKR response in rlbp1b knockout larvae. p.R151Q 3 biological replicates n ≥ 12 larvae per group (p < 0.0001, two tailed t‐test). p.A237V 3 biological replicates n ≥ 11 larvae per group (p < 0.0047, Welch's t‐test). (F) Schematic of expression vectors used to create zebrafish lines expressing EGFP‐tagged rlbp1b used in this study. (G) F0 rescue of the dim light OKR phenotype using the tg(rpe65a:eGFP‐rlbp1b) transgene in the rlbp1b knockout larvae. Three biological replicates n = 24 rlbp1b −/− larvae per group (p = 0.0086, two tailed t‐test). (H) Rescue of the dim light OKR phenotype in F1 larvae stably expressing tg(rpe65a:eGFP‐rlbp1b). Three biological replicates n = 27 rlbp1b −/− larvae per group (p < 0.0001, two tailed t‐test). (I) Stable expression of the tg(rpe65a:eGFP‐rlbp1bp.R151Q) fails to rescue the dim light OKR phenotype in rlbp1b knockout larvae. No significant difference between non‐transgenic larvae and larvae with the tg(rpe65a:eGFP‐rlbp1bp.R151Q) transgene. Significant difference between larvae with the wildtype vs. p.R151Q mutant transgene (p < 0.0053, Kruskal–Wallis test). Three biological replicates n = 25 rlbp1b −/− larvae per group. All graphs show mean ± standard deviation.
Fig 3: Zebrafish rlbp1a is expressed in the Müller glia and rlbp1b is expressed in the RPE. (A) Schematic diagram of the role of CRALBP in the visual cycle. Created using biorender.com. (B) tSNE clustering of scRNAseq data from [33]. (C) Expression of rlbp1a from the [33] dataset showing expression in Müller glia (red dots). (D) Expression of rlbp1b from the [33] dataset showing expression in the RPE (red dots). (E) UMAP projection of gene expression from the Daniocell database (daniocell.nichd.nih.gov). (F) UMAP projection of rlbp1a expression from Daniocell showing expression in Müller glia (blue to red colouration). (G) UMAP projection of rlbp1b expression showing expression in the RPE (blue to red colouration). (H) Time course of rlbp1a expression in the RPE and Müller glia during development from Daniocell. (I) Time course of rlbp1b expression in the RPE and Müller glia during development from Daniocell.
Fig 4: Human RLBP1 represents a good target for a humanized zebrafish approach. (A) Schematic of the humanized zebrafish lines approach used in this study. (B) Alignment of Cralbp protein sequences using the TCoffee alignment tool. Domains mapped to these sequences manually from the NCBI conserved domain search tool. Pathogenic mutations in CRALBP are taken from [6]. (C) Structural model of human CRALBP and zebrafish paralogs created using the SWISS‐MODEL web tool. (D) Structural model of human CRALBP and missense variants p.R151Q and p.A237V using the SWISS‐MODEL web tool.
Supplier Page from DNASU for RLBP1 (Homo sapiens) in pDONR221 (Gateway donor/master vector)