Yap and Taz regulate retinal pigment epithelial cell fate

Abstract

The optic vesicle comprises a pool of bi-potential progenitor cells from which the retinal pigment epithelium (RPE) and neural retina fates segregate during ocular morphogenesis. Several transcription factors and signaling pathways have been shown to be important for RPE maintenance and differentiation, but an understanding of the initial fate specification and determination of this ocular cell type is lacking. We show that Yap/Taz-Tead activity is necessary and sufficient for optic vesicle progenitors to adopt RPE identity in zebrafish. A Tead-responsive transgene is expressed within the domain of the optic cup from which RPE arises, and Yap immunoreactivity localizes to the nuclei of prospective RPE cells. yap (yap1) mutants lack a subset of RPE cells and/or exhibit coloboma. Loss of RPE in yap mutants is exacerbated in combination with taz (wwtr1) mutant alleles such that, when Yap and Taz are both absent, optic vesicle progenitor cells completely lose their ability to form RPE. The mechanism of Yap-dependent RPE cell type determination is reliant on both nuclear localization of Yap and interaction with a Tead co-factor. In contrast to loss of Yap and Taz, overexpression of either protein within optic vesicle progenitors leads to ectopic pigmentation in a dosage-dependent manner. Overall, this study identifies Yap and Taz as key early regulators of RPE genesis and provides a mechanistic framework for understanding the congenital ocular defects of Sveinsson's chorioretinal atrophy and congenital retinal coloboma.

Keywords: Choroid fissure; Coloboma; Directed differentiation of stem cells; Eye development; Hippo signaling; Ocular morphogenesis; Sveinsson's chorioretinal atrophy; Tfec; Zebrafish.

Fig. 1.

A Yap/Taz-Tead reporter transgene is dynamically expressed during optic cup morphogenesis and yap^−/− mutants exhibit RPE defects. (A-D′) Images of live zebrafish from 14-24 hpf showing optic cup development and 4xGTIIC:d2GFP transgene expression (green). Arrows indicate cells that are expressing the transgene while undergoing morphogenesis. (E) Schematic of wild-type and mutant Yap and Taz. Yap S54 is an essential residue for Tead binding and S87 is phosphorylated by Lats leading to cytoplasmic retention. The Yap c.158_161del and Taz c.156_160del mutants contain frameshifts resulting in early stop codons. TEAD BD, Tead transcription factor binding domain (light blue); WW, dual tryptophan motif (green); TAD, transactivation domain (fushia); PDZ domain (dark blue). (F,G) yap^−/−; taz^−/− embryos arrest by 18 hpf with multiple defects. (H-J″) Live embryos (H-J′) and sections (H″,I″,J″) of yap^−/−; taz^+/+ (I-I″) and yap^−/−; taz^+/− (J-J″) showing RPE defects and additional NR defects in yap^−/−; taz^+/− mutants (J′) compared with control (H-H″). Boxed areas indicate locations of TEM analysis. (K-L′) Transmission electron microscopy analysis showing areas of normal RPE development (L′) and areas devoid of RPE (L) in yap^−/− eyes. Asterisk indicates the presence of primary cilia on neuroepithelial cells. L, lens; OV, optic vesicle; NR, neural retina; RPE, retinal pigment epithelium; SE, surface ectoderm; POM, periocular mesenchyme; NP, neuropil; PhRP, photoreceptor progenitors.

Fig. 2.

yap^nl13/nl13 mutants exhibit coloboma. (A) Splice site variants elicited by the yap^nl13 allele. The mutation results in aberrant splicing and an early stop codon in the transactivation domain for all described products. (B) Protein schematics for wild-type and predicted mutant yap^nl13 allele variants. (C) Western blot showing the loss of full-length Yap in yap^nl13/nl13 mutants. A smaller protein product is detectable at higher levels in the mutant (∼40 kDa). (D-D″,G-G″) 3 dpf wild-type and yap^nl13/nl13 embryos showing coloboma (arrows) in the absence of other overt phenotypes. Plastic sections of wild-type (D″) and mutant (G″) eyes show the coloboma (arrow) phenotype and RPE deficits that are sometimes observed in the ventral retina. (E,F,H,I) Sections showing Yap and Taz proteins (green) in wild-type and yap^nl13/nl13 mutant eyes. Red counterstain (TO-PRO3) shows nuclei.

Fig. 3.

yap mRNA and Yap protein levels are decreased and Taz protein increased in yap^−/− embryos. (A-B‴) Yap immunoreactivity in wild-type and yap^−/− eyes at 28 hpf. Yap protein is present in flattened RPE nuclei (arrows) and periocular mesenchyme (POM) in yap^+/− embryos, whereas nuclear Yap staining is absent in the yap^−/− mutant. (C) qRT-PCR analysis of whole embryos at 32 hpf showing a decrease in yap (3-fold, *P=0.0002) and taz (1.5-fold, *P=0.0270) mRNA in yap^−/− mutants. Dotted line indicates normalized expression levels of yap and taz in wild-type embryos. An unpaired t-test was performed and statistical significance determined using the Holm-Sidak method. Error bars represent s.e.m. (D) Western blot showing Taz protein (∼52 kDa) in wild-type and its absence in taz^−/− adult heart tissue. (E-G‴) Taz immunoreactivity in wild-type, taz^−/− and yap^−/− embryos at 28 hpf. (H) Western blot of Taz protein from 2 dpf wild-type (n=20) and yap^−/− mutant (n=20) whole embryos.

Fig. 4.

Yap and Tead1a zebrafish protein interactions are conserved. (A) Schematic of the zebrafish Yap and Tead binding domain (BD) interaction sites. (B) Immunoprecipitation of zebrafish Yap and Tead1a wild-type and Tead-binding-deficient Yap (Yap S54A) and Yap-binding-deficient Tead1a (Tead1a Y417H) isoforms. All the mutated protein variants lose the ability to interact, in contrast to the wild-type proteins. Immunoprecipitation (IP) was with an anti-Flag antibody. (C-E″) Whole eyes (C-D′) and sections (E-E″) showing RPE loss surrounding the optic nerve head after overexpression of Tead1a Y417H. Photoreceptors are red owing to late expression of the rx3:Gal4 driver in these cells. Arrows indicate areas lacking RPE.

Fig. 5.

Tead-binding-deficient yap^ΔTB/ΔTB mutants lack RPE but maintain yap mRNA and Yap protein levels. (A) The Tead-binding-deficient yap^ΔTB/ΔTB zebrafish mutant. (B-D) 48 hpf whole eyes showing that yap^−/− and yap^ΔTB/ΔTB mutants lose a subset of RPE cells. Dashed lines indicate the border of the eye. (E) qRT-PCR analysis of whole embryos at 32 hpf revealing a decrease in yap (1.6-fold, *P=0.0052) and taz (1.15-fold, *P=0.0038) mRNA in yap^ΔTB/ΔTB mutants. Dotted line indicates the normalized expression levels of yap and taz in wild-type embryos. An unpaired t-test was performed and statistical significance was determined using the Holm-Sidak method. Error bars indicate s.e.m. (F-G‴) Yap protein expression in yap^+/ΔTB (F-F‴) and yap^ΔTB/ΔTB (G-G‴) at 28 hpf. Yap protein is present in flattened RPE nuclei (arrows) and periocular cells in yap^+/ΔTB and yap^ΔTB/ΔTB embryos.

Fig. 6.

Cell death and proliferation are normal in yap^−/− eyes. (A) The numbers of dying cells identified by Acridine Orange staining do not differ in yap^−/− mutants at 14 (P=0.8465), 18 (P=0.6542) or 24 (P=0.2558) hpf as compared with wild type. Numbers of eyes analyzed: 14 hpf, wt n=31, yap^−/− n=5; 18 hpf, wt n=39, yap^−/− n=10; 24 hpf, wt n=24, yap^−/− n=16. (B) Eye field mitotic cell counts do not differ between yap^+/? (wt) and yap^−/− at 14 (P=0.9205), 18 (P=0.4329) and 24 (P=0.2222) hpf. Numbers of eyes analyzed: 14 hpf, wt n=38, yap^−/− n=6; 18 hpf, wt n=38, yap^−/− n=6; 24 hpf, wt n=36, yap^−/− n=6. (C-E‴) Time-course of expression of tfec:eGFP in prospective RPE cells. (F) Mitotic cell counts of tfec:eGFP⁺ cells do not differ between wild-type and yap^−/− eyes at 14 hpf (P=0.5408). Numbers of eyes analyzed: wt n=18, yap^−/− n=6. P-values were obtained using an unpaired t-test with equal s.d. Error bars indicate s.e.m. Wild type included full RPE coverage, whereas yap^−/− showed some RPE loss.

Fig. 7.

yap^−/−; taz^−/− transplanted cells do not contribute to the RPE. (A) Methods used to analyze the contribution of transplanted cells to the RPE and NR. (B-C″) Examples of transplanted wild-type (B-B″) and yap^−/−; taz^−/− (C-C″) cells in 48 hpf eyes. The arrow in B indicates a pigmented transplanted cell forming RPE in the albino host; other black/white arrows indicate H2a-GFP⁺ clones in NR.

Fig. 8.

Overexpression of gain- and loss-of-function yap transgenes alters RPE and NR cell fate. (A-D′) Images showing that overexpression of wild-type Yap (B,B′), constitutively active Yap (Yap S87A) (C,C′) and Taz (D,D′) induces ectopic pigmentation and retinal disorganization at 5 dpf, as compared with the control (A,A′). Arrows indicate the presence of ectopic pigment cells. (E-G) Mosaic overexpression of dominant-negative forms of Yap (F) and Taz (G) results in ectopic loss of RPE cells at 48 hpf, as compared with the control (E). Arrows indicate areas devoid of RPE cells. (H) Model of zebrafish RPE development. Bipotent progenitor cells assume either an RPE or NR fate based on Yap/Taz activity. In the absence of yap (right hemisphere of the eye cup), progenitors contributing to the central retina cannot form RPE. Those progenitors that contribute to the peripheral eye cup can upregulate Taz during their more lengthy migration and therefore form RPE.