In Vivo Analysis of Disease-Associated Point Mutations Unveils Profound Differences in mRNA Splicing of Peripherin-2 in Rod and Cone Photoreceptors

Abstract

Point mutations in peripherin-2 (PRPH2) are associated with severe retinal degenerative disorders affecting rod and/or cone photoreceptors. Various disease-causing mutations have been identified, but the exact contribution of a given mutation to the clinical phenotype remains unclear. Exonic point mutations are usually assumed to alter single amino acids, thereby influencing specific protein characteristics; however, they can also affect mRNA splicing. To examine the effects of distinct PRPH2 point mutations on mRNA splicing and protein expression in vivo, we designed PRPH2 minigenes containing the three coding exons and relevant intronic regions of human PRPH2. Minigenes carrying wild type PRPH2 or PRPH2 exon 2 mutations associated with rod or cone disorders were expressed in murine photoreceptors using recombinant adeno-associated virus (rAAV) vectors. We detect three PRPH2 splice isoforms in rods and cones: correctly spliced, intron 1 retention, and unspliced. In addition, we show that only the correctly spliced isoform results in detectable protein expression. Surprisingly, compared to rods, differential splicing leads to lower expression of correctly spliced and higher expression of unspliced PRPH2 in cones. These results were confirmed in qRT-PCR experiments from FAC-sorted murine rods and cones. Strikingly, three out of five cone disease-causing PRPH2 mutations profoundly enhanced correct splicing of PRPH2, which correlated with strong upregulation of mutant PRPH2 protein expression in cones. By contrast, four out of six PRPH2 mutants associated with rod disorders gave rise to a reduced PRPH2 protein expression via different mechanisms. These mechanisms include aberrant mRNA splicing, protein mislocalization, and protein degradation. Our data suggest that upregulation of PRPH2 levels in combination with defects in the PRPH2 function caused by the mutation might be an important mechanism leading to cone degeneration. By contrast, the pathology of rod-specific PRPH2 mutations is rather characterized by PRPH2 downregulation and impaired protein localization.

Fig 1. Design and in vivo expression of PRPH2 minigenes.

(A) Exon-intron structure of native human PRPH2 (upper panel) and derived PRPH2 minigene used in this study (lower panel). In the minigene, intron 1 and intron 2 were largely deleted with the exception of sequences flanking exons 1–3 as indicated. (B) PRPH2 minigene constructs containing different promoters used for the analysis in rods (rP-mg; top), and cones (cP-mg; bottom). hRHO, human rhodopsin promoter; mSWS, murine S-opsin promoter. To allow antibody-free visualization of minigene-derived PRPH2, a citrine tag was fused to exon 1 (N-terminus) of PRPH2. (C) Topology of the correctly spliced citrine-tagged PRPH2. N- and C-termini face the intracellular side of photoreceptors. By contrast, the D2 loop is located on the intradiskal (rods) or extracellular (cones) side (cf. S1 Fig). (D and E) Immunohistology of murine retinas injected on P14 with rP-mg (D) and cP-mg (E), respectively. Retinas were harvested three weeks post injection. Scale bar represents 20 μm. CNGB1a (B1a) and M-opsin (M-ops) antibodies were employed to label rod and cone outer segments, respectively.

Fig 2. Splice analysis of PRPH2 WT and mutant minigenes in rods and cones.

(A) Representative RT-PCR from cDNA generated from total RNA three weeks post injection from retinas injected with wild-type and mutant rP-mg (left) or cP-mg (right) on P14. Ctrl, control containing the cDNA from non-transduced retina. The single bands of the relevant splice products are numbered (1–4) and highlighted by arrowheads. (B) Schematic representation of the detected splice variants using primers binding to the 3’-end of citrine and to the 5’-end of exon 3 as indicated by the arrows. The numbers of the constructs correspond to the bands marked in (A). (C-E) Semi-quantitative analysis of the relative intensities of the unspliced (C), intron 1 retention (D), and correctly spliced (E) PRPH2 transcripts. For each PRPH2 minigene, the mean percentage of the intensities of these three variants relative to the total intensity (given as sum of the single intensities) was calculated from five RT-PCR analyses conducted with a variable number of cycles (25–27 for rods and 30–32 for cones, respectively). Significance test of the rod or cone mutants to the corresponding wild type (one-way ANOVA followed by Dunett’s test) was performed for rP-mg and cP-mg, respectively. All data were shown as mean values and the error bars represent the standard error of the mean (SEM). *, p< 0.05; **, p< 0.01; ***, p< 0.001. DS, splice donor site.

Fig 3. Quantification of PRPH2 splice isoforms in native human and murine photoreceptors.

(A) True-to-scale representation of the exonic and intronic regions of human PRPH2 (upper panel) and mouse Prph2 (lower panel). The binding positions of primers used for qRT-PCR are indicated by arrows. For correctly spliced human PRPH2 or murine Prph2 (hP-cs or mP-cs), the P-cs_F and P-cs_R primer combination was used. For detection of the unspliced human (hP-us) or murine (mP-us) variant, the P-us_F and P-us_R primer combination was applied. (B) qRT-PCR from pooled human (gray boxes) or mouse (black boxes) total retinal RNA isolated from two (human) or four (mouse) biological samples. The single values are as follows: hP-cs, 10.60 ± 0.67; hP-us, 0.21 ± 0.03; mP-cs, 5.60 ± 0.29; mP-us, 0.13 ± 0.03. All data are given as mean values and error bars represent the SEM. Three technical replicates were conducted for each primer combination and the expression was normalized to the housekeeper aminolevulinic acid synthase (ALAS). (C) Relative ratios given as mean values ± SEM of the single unspliced transcripts to the corresponding correctly spliced isoform from human (0.10 ± 0.01) or mouse (0.13 ± 0.02) retina. Significance test was performed with the two-tailed t-test (p = 0.26). (D) qRT-PCR analysis from sorted murine rods (red boxes) and cones (black boxes), respectively. For qRT-PCR, cDNA resulting from pooled rods sorted from six animals (yielding 100.000 cells) and pooled cones sorted from four animals (yielding 27.000 cells) were used. cDNA synthesis was performed using identical total RNA concentration for rods and cones, respectively (50 ng each). For purity assessment of FAC-sorted rods and cones, primers specific for murine rhodopsin (Rho) and M-opsin (Opn1mw) were used. For detection of Prph2 splice isoforms, the same primer combinations were applied as indicated in Fig 3A. Three technical replicates were conducted for each primer combination and the expression was normalized to the housekeeper aminolevulinic acid synthase (Alas). All data are given as mean values and error bars represent the SEM. The single values for FAC-sorted rods are as follows: Rho, 146.90 ± 10.24; Opn1mw, 1.25 ± 0.16; P-cs, 16.67 ± 0.71; P-us, ± 0.95 ± 0.07. Expression in sorted cones yielded following values: Rho, 5.27 ± 0.14; Opn1mw, 6.38 ± 0.13; P-cs, 5.45 ± 0.44; P-us, ± 2.71 ± 0.19. P-cs rods vs cones: p = 0.0002; P-us rods vs cones: p = 0.001. (E) Relative ratios of the single unspliced transcripts to the corresponding correctly spliced isoform from FAC-sorted rods or cones. Significance analysis was performed with the two-tailed t-test (p = 0.0003).

Fig 4. Impaired targeting of truncated PRPH2.

Immunohistology of transduced murine retinas showing rod- (B) and cone-specific (C) expression of a truncated version of PRPH2. (A) Truncated PRPH2 contains only exon 1 and a downstream stop codon (indicated by “X”) mimicking translation from unspliced PRPH2 mRNA and PRPH2 mRNA with intron 1 retention. Staining for B1a and M-ops was used to label rod and cone photoreceptors, respectively. Truncated PRPH2 is not transported to outer segments and is almost exclusively present in inner segments and somata of photoreceptors. Scale bar represents 20 μm.

Fig 5. Retinal expression of PRPH2 mutants linked to cone diseases.

(A-E) Immunohistological analysis of retinas transduced with the PRPH2 minigenes carrying single point mutations under the control of the mSWS promoter as indicated. Scale bar represents 20 μm. (F) Western blot analysis from membrane preparations of murine retinas transduced with the PRPH2 minigenes shown in A-E. Western blotting was conducted using four pooled retinas from four animals injected on P14. All retinas were collected three weeks post injection. Ctrl, protein lysates from non-injected control retinas. PRPH2 was detected by an anti-GFP antibody that recognized the citrine tag. As loading control, an antibody against the murine alpha subunit of ATPase was used (anti-ATPase). (G) Semi-quantitative analysis of the results shown in (F). For quantification, three technical replicates were conducted and PRPH2 expression was normalized to the ATPase expression. All data are given as mean values ± SEM. Statistical analysis was performed using one-way ANOVA followed by the Dunett’s test. *, p< 0.05; **, p< 0.01; ***, p< 0.001. n.s., not significant.

Fig 6. Protein expression of PRPH2 mutants linked to adRP.

(A-F) Immunohistology of retinas transduced with PRPH2 minigenes containing single point mutations as indicated under the control of the hRHO promoter. Scale bar represents 20 μm. (G) Western blot analysis from membrane preparations of four pooled murine retinas from four animals transduced with the PRPH2 minigenes shown in (A-F) on P14. All retinas were collected three weeks post injection. The arrowhead indicates a degradation band detected at 42 kDa. Ctrl, protein lysates from non-injected control retinas. (H) Semi-quantitative analysis of the results shown in (G). For quantification, three technical replicates were conducted and PRPH2 expression was normalized to the ATPase expression. All data are given as mean values and error bars represent the SEM. Statistical analysis was performed using one-way ANOVA followed by the Dunett’s test. *, p< 0.05; **, p< 0.01; ***, p< 0.001. n.s., not significant.