Mutant prominin 1 found in patients with macular degeneration disrupts photoreceptor disk morphogenesis in mice

Abstract

Familial macular degeneration is a clinically and genetically heterogeneous group of disorders characterized by progressive central vision loss. Here we show that an R373C missense mutation in the prominin 1 gene (PROM1) causes 3 forms of autosomal-dominant macular degeneration. In transgenic mice expressing R373C mutant human PROM1, both mutant and endogenous PROM1 were found throughout the layers of the photoreceptors, rather than at the base of the photoreceptor outer segments, where PROM1 is normally localized. Moreover, the outer segment disk membranes were greatly overgrown and misoriented, indicating defective disk morphogenesis. Immunoprecipitation studies showed that PROM1 interacted with protocadherin 21 (PCDH21), a photoreceptor-specific cadherin, and with actin filaments, both of which play critical roles in disk membrane morphogenesis. Collectively, our results identify what we believe to be a novel complex involved in photoreceptor disk morphogenesis and indicate a possible role for PROM1 and PCDH21 in macular degeneration.

Figure 1. Retinal degeneration as a consequence of mutant PROM1.

(A and B) Clinical features of autosomal-dominant macular dystrophy and segregation of a PROM1 mutation. (A) Left: Fundus photograph from a STGD4 patient with visual acuity of 20/200, showing an area of macular atrophy with surrounding yellow flecks (white arrows). Right: Fundus photograph from an MCDR2 patient with a visual acuity of 20/80, showing bull’s-eye maculopathy (white arrows). (B) Sequencing traces demonstrating an 1117 C>T transition in exon 10 giving rise to the missense R373C substitution in both families. (C) Fundus photographs from representative 12-month-old PTW20 (left), 4-month-old PMT14 (middle), and 13-month-old PMT14 (right) mice. Yellow arrows denote yellow deposits and small lesions scattered throughout the central fundus. Black arrows denote large coalescing deposits and atrophic lesions in the fundus. (D) Progressive photoreceptor loss, as determined by the number of nuclei in the ONL, in PMT3 and PMT14 mice compared with normal C57BL/6 (control) and PWT20 mice. Photoreceptor cell nuclei were counted in transgenic mice between 4 and 40 weeks of age. Error bars, which in most cases are too small to be visible, represent ± SEM. The rate of photoreceptor cell loss was greater in the PMT14 line, which expressed higher levels of mutant PROM1 than did PMT3 mice (see Figure 3).

Figure 2. Electron microscopy of PROM1 transgenic mouse rod photoreceptors.

(A and B) Overgrown and abnormally oriented disk membranes (arrows) in PMT3 retinas were visible at 16 days (A) and 4 months (B) of age. Arrowheads indicate vesiculated membrane. Higher-magnification view of the boxed region in B is shown in the inset. (C) Electron micrograph showing abnormalities in the RPE of PMT3 mice at 4 months of age. Accumulation of lipofuscin-like material (arrows) was evident. (D) Electron micrograph of a WT retina, showing normal OS disk membranes and adjacent RPE. Scale bars: 1 μm (A, B, and D); 300 nm (C); 30 nm (B, inset).

Figure 3. Immunohistochemical localization of human PROM1 and endogenous mouse PROM1 in PROM1 transgenic retinas at 1 month of age.

(A–L) Localization of transgenic human PROM1 (red) and native mouse PROM1 (green). (A–C) C57BL/6 mice showed endogenous mouse PROM1 signal alone at the base of the OSs. (D–F) In PWT20 mice, human PROM1 and endogenous mouse PROM1 were normally localized to the base of the OSs. (G–L) In PMT3 and PMT14 retinas, mutant PROM1 was mislocalized throughout the photoreceptors from the ONL to the OSs, whereas endogenous mouse PROM1 signal was more restricted to the ISs and OSs. Both endogenous mouse PROM1 and mutant PROM1 showed strong immunolabeling in the myoid region of the ISs. (M–R) PROM1 (red) and PCDH21 (green) protein immunolocalization in WT and mutant PROM1 transgenic retinas. (M–O) In 1-month-old PWT20 mice, PCDH21 and PROM1 colocalized at the base of photoreceptor OSs. (P–R) In PMT14 mice, PCDH21 was mislocalized, as was PROM1, to the ONL. (S–U) WT PROM1 immunolocalization in Pcdh21^–/– mice. WT PROM1 immunolabeling (S) was evident throughout OSs. PROM1 colocalized with the OS marker CNGCA1 (T) and not the IS marker, Na/K-ATPase (U). Note that CNGCA1 and Na/K-ATPase signals were correctly located within OS and IS compartments, similar to the normal immunolabeling pattern in the R373C mutant PROM1 transgenic mice. INL, inner nuclear layer; OPL, outer plexiform layer; IPL, inner plexiform layer; GC, ganglion cells. Scale bars: 20 μm.

Figure 4. Coimmunoprecipitation of PROM1 and PCDH21 and proteolytic cleavage of PCDH21.

(A) HEK293 cells were cotransfected with PCDH21 and either WT or mutant PROM1. Immunoprecipitation was performed with anti-PCDH21, followed by anti-PROM1 immunodetection on a Western blot. Immunolabeling of WT (lane 1) and mutant (lane 3) PROM1 in the transfected cell lysates (anti-PROM1 Western blot) served as positive controls. Also shown is immunolabeling of WT (lane 2) and mutant (lane 4) PROM1 in the PCDH21 immunoprecipitates. (B) Reciprocal immunoprecipitation with PROM1 antibody after HEK293 cells were cotransfected with PCDH21-Myc and WT (lane 2) or mutant (lane 3) PROM1. Lane 1 shows WT PCDH21-Myc–transfected cell lysate (anti-Myc Western blot) as a positive control. (C) Immunoprecipitation of retinal lysates. Lysates from WT (lane 2) and mutant (lane 3) PROM1 transgenic mice were immunoprecipitated with anti-PCDH21, followed by anti-human PROM1 detection on a Western blot. Lane 1 is a PROM1 protein positive control, showing PROM1 immunolabeling of retinal lysate. (A–C) Results of negative control experiments performed using nonspecific IgG were negative. (D) Proteolytic processing of PCDH21 in PROM1 transgenic mice. Full-length (FL) and proteolytic C-terminal (CT) PCDH21 fragments were compared among PMT3 (lane 1), C57BL/6 (lane 2), PWT31 (lane 3), and PWT20 (lane 4) mice. Total retinal protein lysate (20 μg) was probed with the PCDH21 C terminus antibody; β-actin was used as a loading control. PMT3 retinas showed substantially more full-length PCDH21 and less cleaved fragment than did PWT20, PWT31 and C57BL/6 retinas. Lanes were run on the same gel but were noncontiguous.

Figure 5. Coimmunoprecipitation of PROM1 and β-actin.

(A) HEK293 cells were transfected with either WT or mutant (MT) PROM1. Whole cell lysates with equal amounts of input WT or mutant PROM1 were immunoprecipitated with PROM1 polyclonal antibody, followed by anti–β-actin immunodetection on Western blots (WB). (B) Quantification of PROM1 and β-actin interaction by densitometry. Values (mean ± SD) denote the amount of β-actin, immunoprecipitated with PROM1 antibodies from cells, transfected with WT or mutant PROM1. Each experiment was done in triplicate, and 3 independent transfection experiments were performed for each PROM1 construct. Significance was examined using independent sample t test.