Knockout of RP2 decreases GRK1 and rod transducin subunits and leads to photoreceptor degeneration in zebrafish

Abstract

Retinitis pigmentosa (RP) affects about 1.8 million individuals worldwide. X-linked retinitis pigmentosa (XLRP) is one of the most severe forms of RP. Nearly 85% of XLRP cases are caused by mutations in the X-linked retinitis pigmentosa 2 (RP2) and RPGR. RP2 has been considered to be a GTPase activator protein for ARL3 and to play a role in the traffic of ciliary proteins. The mechanism of how RP2 mutations cause RP is still unclear. In this study, we generated an RP2 knockout zebrafish line using transcription activator-like effector nuclease technology. Progressive retinal degeneration could be observed in the mutant zebrafish. The degeneration of rods' outer segments (OSs) is predominant, followed by the degeneration of cones' OS. These phenotypes are similar to the characteristics of RP2 patients, and also partly consistent with the phenotypes of RP2 knockout mice and morpholino-mediated RP2 knockdown zebrafish. For the first time, we found RP2 deletion leads to decreased protein levels and abnormal retinal localizations of GRK1 and rod transducin subunits (GNAT1 and GNB1) in zebrafish. Furthermore, the distribution of the total farnesylated proteins in zebrafish retina is also affected by RP2 ablation. These molecular alterations observed in the RP2 knockout zebrafish might probably be responsible for the gradual loss of the photoreceptors' OSs. Our work identified the progression of retinal degeneration in RP2 knockout zebrafish, provided a foundation for revealing the pathogenesis of RP caused by RP2 mutations, and would help to develop potential therapeutics against RP in further studies.

Introduction

Retinitis pigmentosa (RP, OMIM #268000) is an inherited and progressive retinal degeneration disease caused by gradual dysfunction and apoptosis of photoreceptors. It has a worldwide prevalence ranging from 1/3000 to 1/7000 with great genetic heterogeneity (1–4). The manifestations of RP patients are highly variable, usually including night blindness (early stages), peripheral retinal pigmentation and vision loss, central vision loss and retinal vessels attenuation (late stages). X-linked retinitis pigmentosa (XLRP) is one of the most severe forms of RP, characterized by early-onset and rapid progression of visual impairment. Mutations in the X-linked retinitis pigmentosa 2 (RP2, OMIM *300757) gene account for 7–18% of XLRP cases (5,6).

RP2 was identified in 1998 (7). It consists of five exons encoding a ubiquitously expressed 350-amino acid polypeptide, which contains an N-terminal tubulin folding cofactor C-like (TBCC) domain and a C-terminal nucleoside diphosphate kinase-like (NDPK) domain. So far, more than 70 RP2 mutations have been reported and over half of them are located in exon 2 (8). Currently, the mechanism of how RP2 mutations cause retinal degeneration is still not fully understood.

In human retinas, RP2 associates with the plasma membrane of photoreceptors, including the outer segments (OSs), inner segments, cell bodies and synapses (9). The membrane localization of RP2 relies on the myristoylation of Gly2 and palmitoylation of Cys3 at the N-terminal. RP2 has also been reported to locate to the cilium and the Golgi apparatus in cultured cell lines (10–12). Several proteins have been identified to interact with RP2, including tubulin (13), ADP-ribosylation factor-like 3 (ARL3) (14), polycystin 2 (11), Importin β2 (10), Gβ1 (β subunit of rod transducin) (15) and N-ethylmaleimide sensitive factor (NSF) (16). Among them, the small GTPase ARL3 is the best-studied RP2 relevant protein. Structural analysis and biochemical research revealed that RP2 accelerates the GTPase activity of ARL3, which dissociates the ARL3-GTP-UNC119 complex after ARL3-GTP releases myristoylated cargos from UNC119 (14,17–19). The UNC119 homologous protein, PDE6D, which is another ARL3 effector, can bind prenylated proteins and plays a role in maintaining the GRK1 and PDE6 catalytic subunits in mouse retinas (20–22). The particular correlation between RP2 and PDE6D has not been confirmed experimentally until now. Deletion of ARL3 (23), UNC119 (24) and PDE6D (22) all leads to retinopathy in mice. Furthermore, RP2 has been reported to facilitate the membrane association of GNB1 in ARPE19 cells (15). Up to now, two mouse models of RP2 have been described in 2013 (25) and recently in 2014 (26). Progressive retinal degeneration with a cone predominantly affected pattern was identified in both RP2 defective models without any obvious physical defect.

In recent years, zebrafish is becoming a useful animal model due to the convenience of genetic manipulation and the high rate of spawning. Additionally, the similar structure of retina and the cone-dominant vision of zebrafish (similar to humans) makes it an important model for human retinal disorders (27,28). In zebrafish, there is only one orthologue of RP2, containing 6 exons coding 376 amino acids. Morpholino-mediated knockdown of RP2 in zebrafish leads to severe early-onset retinal dysplasia (29,30) and pronephric cysts (11), with body curvature, left–right symmetry defects and heart looping defects, implying the dysfunction of cilium. The high amino acid identity (67%; Supplementary Material, Fig. S1) and the rescue data in zebrafish using human RP2 mRNA (29,30) suggest that the function of RP2 may be conserved between humans and zebrafish.

To explore the phenotype of RP2 defective zebrafish and the pathogenesis of XLRP caused by RP2 mutations, we generated an RP2 knockout zebrafish line in this study by transcription activator-like effector nuclease (TALEN) technology. Progressive rod degeneration followed by cone degeneration could be observed in RP2 knockout zebrafish, as well as a decline in GRK1, GNAT1 and GNB1 protein levels. Our study provides a foundation for revealing the causes of RP2-associated RP and developing potential therapeutics against RP.

Results

Generation of RP2 knockout zebrafish with TALEN technology

The zebrafish RP2 cDNA (NM_213446) and genomic sequences were downloaded from the NCBI database. The TBCC domain of RP2 encoded by exon 2 in zebrafish is highly conserved between human and zebrafish with 80% identity (Fig. 1A and Supplementary Material, Fig. S1). Exon 2 is also a mutational hotspot in RP2 patients (8,31). Therefore, we designed the target site of TAL effectors (TALEs) in exon 2. The details of RP2 TALEs used in this study are shown in Figure 1A. The TALEN mRNAs were obtained by in vitro transcription and injected into zebrafish embryos. The effective groups of injected embryos were raised to sexual maturity (F0 zebrafish).

Figure 1.

Generation of the RP2 knockout zebrafish. (A) The six exons of zebrafish RP2 and their corresponding positions in protein level were shown. The TALE binding sequences were underlined. TALE-L and TALE-R, the left and right arms of the TALEs; the AvaII restriction site in the spacer region is used for mutation detection. (B) Sequencing of the c.359_363delCGGTC (del5) RP2 mutation in homozygous zebrafish. The 5-bp deletion was indicated with a box. (C) RP2 protein levels of WT and RP2^null zebrafish at 10 dpf revealed by western blot using the anti-human RP2 antibody. Molecular mass marker (M) was shown in kDa. GAPDH, loading control; arrow, RP2 bands; NS, non-specific bands. (D) RP2 mRNA levels in 2-month-old WT and RP2^null zebrafish eyes detected by quantitative PCR. Actb1 was served as endogenous control. The result was shown as mean ± SD. **P < 0.01. (E) Immunostaining of the adult WT and del5 mutant retinas using the antibody against zebrafish RP2. The middle panel shows the result of negative control, which used RP2 peptide to block the antibody. RPE, retinal pigment epithelium; OS, outer segment; IS, inner segment; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer, GCL, ganglion cell layer. Scale bar: 20 µm.

Mutant screening was performed in three generations of zebrafish. We identified five mutations (Supplementary Material, Table S1), including four truncation mutations and one in-frame deletion. The percentage of mutant zebrafish in examined F1 animals was 35% (13 of 37), indicating a high efficiency of TALEN-induced mutagenesis. The c.359_363delCGGTC (p.Pro120Glnfs*10) mutation (named del5) was the most frequent one, and was predicted to abolish the function of RP2. Homozygous del5 zebrafish were identified. Their offsprings (F3 zebrafish) were validated by genomic DNA sequencing (Fig. 1B).

The RP2 protein could not be detected in del5 mutant homozygotes using the antibody against human RP2 that could recognize the wild-type zebrafish RP2 (Fig. 1C). The mRNA levels of RP2 in 2-month-old del5 zebrafish eyes were reduced to nearly 40% compared with those of WT controls (Fig. 1D), probably due to the nonsense-meditated mRNA decay. Because the antibody didn't work in the immunofluorescence assay of zebrafish retinal sections, we generated a rabbit anti-zebrafish RP2 antibody and tested the antibody by western blot and immunostaining in zebrafish. The western blot result was consistent with that of anti-human RP2 antibody (Supplementary Material, Fig. S2A). Co-stained with zpr1 (marker of the double cones' cell bodies) and zpr3 (marker of the rod's OSs), antibodies on 4-month-old WT zebrafish retinal sections revealed that RP2 partly co-localized with zpr1 and surrounded the signal of zpr3 (Supplementary Material, Fig. S2B), indicating RP2 is expressed in both rod and cone photoreceptor cells with mainly plasma membrane localization from the synapses to the OSs. In other types of retinal cells, the signal of RP2 was much weaker. Furthermore, we performed immunostaining in adult WT and del5 mutant zebrafish using the anti-zebrafish RP2 antibody. In del5 mutant group or RP2 peptide blocking group, no specific signal could be detected, whereas in WT group, we observed similar retinal localization of RP2 as mentioned earlier (Fig. 1E). These results demonstrate that the knockout of RP2 in this study is effective. In the rest of the research, we considered the homozygous del5 zebrafish as an RP2 null animal model.

Truncation mutations in RP2 have been reported to cause protein degradation or aggregation in insoluble fractions with normal mRNA levels in lymphoblastoid cell lines derived from RP2 patients (32). Recently, Schwarz et al. reported that in patient (R120X) iPSC-derived RPE cells, mutant RP2 protein could not be detected and the mRNA level of RP2 R120X transcript was also decreased (33), which is consistent with the situation of our RP2^null zebrafish. In addition, we did not observe apparent developmental defects such as microphthalmia, body curvature, hydrocephalus and pericardial effusion in RP2^null zebrafish, as have been reported in RP2 knockdown zebrafish previously (11,29,30). These severe developmental defects of RP2 morphants might be caused by the side effects of morpholino injection. The retina laminations of RP2^null zebrafish were also unaffected at the age of 1 month (Supplementary Material, Fig. S3), suggesting that RP2 is seemingly not essential for the early development of zebrafish retina.

RP2^null zebrafish show progressive retinal degeneration with a rod predominantly affected pattern

As reported previously, RP2 mutations cause XLRP in humans (34,35) and rod-cone dystrophy in mice (25,26). We first tested the visual impairment in our RP2^null zebrafish. Electroretinography (ERG) was carried out in 7 dpf (days post-fertilization) zebrafish. The scotopic b-wave amplitudes of RP2^null zebrafish decreased by about 30% when compared with WT controls (Fig. 2), indicating the existence of a mild vision loss at an early age.

Figure 2.

Mild visual impairment in the RP2^null larval detected by ERG analysis. (A) Representative traces of ERG of WT and RP2^null zebrafish at 7 dpf. The arrows indicate the start and the end of light stimulation, respectively. (B) Comparison of b-wave amplitudes between WT (n = 10) and RP2^null (n = 11) zebrafish using two-tailed Student's t-test. The result was shown as mean ± SD. **P < 0.01.

Next, we performed histological analysis to evaluate the morphology of the retinas. At the age of 2 months old, no obvious difference were observed between WT and RP2^null zebrafish, except for the increased pigment granules in the OS layer of the RP2^null zebrafish (Supplementary Material, Fig. S4A). In 7-months-old RP2^null retinas, the abnormal pigmentation was also evident, and the OSs were shortened remarkably when compared with WT control (Supplementary Material, Fig. S4B), suggesting the presence of retinal degeneration.

After that, we wondered whether both the rods and cones were affected and when the abnormities of photoreceptors became obvious. Anti-rhodopsin (4D2) antibody and the peanut lectin (PNA) were used to label the OSs of rods and cones, respectively, in immunofluorescence analysis. At the age of 1 month, the OSs of rods (Fig. 3A, top panel) and cones (Fig. 3B, top panel) exhibited little or no difference between WT and RP2^null zebrafish. In 2-months-old RP2^null retinas, the OSs of rods were mildly shorter than that of WT controls with increased pigment deposition (Fig. 3A, mid-top panel), whereas the OSs of cones were not affected (Fig. 3B, mid-top panel). At 4 months of age, we found remarkable reductions in the length of rods' OS in RP2^null zebrafish with a ∼80% level of WT controls (Fig. 3A, mid-bottom panel and Fig. 3C). Meanwhile, the cones' OS also showed aberrant morphology without decrease in number (Fig. 3B, mid-bottom panel). Furthermore, in 7-months-old RP2^null retinas, the OS degeneration of rods increased (∼70% level of WT controls in OS's length; Fig. 3A, bottom panel and Fig. 3C), and the loss of cones' OS became evident (Fig. 3B, bottom panel, Fig. 3D and Supplementary Material, Fig. S5).

Figure 3.

Progressive retinal degeneration with a pattern of rod mainly affected in the RP2^null zebrafish. Retinal cryosections from WT and RP2^null zebrafish were immunostained with anti-rhodopsin (4D2) antibodies (A) and PNA (B) to label the outer segments of rods and cones at the ages of 1, 2, 4 and 7 months. The white and red lines indicated the thickness of outer segment layer of rods (A) and cones (B), respectively. Scale bars: 20 μm. The statistical data were presented in (C) for rods and (D) for cones. At least, three images of each group were quantified and analysed using two-tailed Student's t-test. The results were shown as mean ± SD. **P < 0.01; *P < 0.05.

To look at the ultrastructure of the OSs of RP2^null zebrafish, we performed transmission electron microscopy assay on the 10-month-old WT and RP2^null zebrafish. The rods' OS almost disappeared in the mutant retina (Fig. 4B and B′) when compared with WT control (Fig. 4A and A′). Among the remaining OSs of the RP2^null retina, some of them showed normal morphology and disc stacking (Fig. 4C and C′), while some of them were shortened and disorganized (Fig. 4D and D′). Taken together, our observations demonstrated that deletion of RP2 leads to progressive retinal degeneration (noticed as early as 2 months old), affects both the rods and cones, but mainly the rods are influenced in early stages in zebrafish.

Figure 4.

Ultrastructural analysis of the WT and RP2^null zebrafish photoreceptors at 10 months of age. (A) Well-maintained rods' outer segments (OSs) from WT retina. (A′) Enlarged image of the box in (A). (B) Nearly complete loss of the rods' OSs in the RP2^null retina. (B′) Enlarged image of the box in (B). In the RP2^null retina, the remaining normal (C) and abnormal (D) OSs were shown. (C′, D′) Enlarged images of the boxes in (C, D). Scale bars: 5 µm in (A, A’ and B, B’); 0.5 µm in (C, C’ and D, D’).

Deletion of RP2 does not affect the retinal localizations of the cone's opsins

Mislocalization of the cone opsins has been reported to be a possible reason of the RP2-associated retinal degeneration (25). In this study, we investigated the retinal localizations of the four cone opsins by immunofluorescence assay. The cone opsins localized to the OSs normally in the 1 and 5.5 months old RP2^null retinas (Supplementary Material, Fig. S6), which is the same as the normal transport of rhodopsin as shown in Figure 3A. Similar result has been found in another RP2 knockout mouse model (26). Thus, the mistrafficking of cone opsins is unlikely to be responsible for the retinal degeneration in our RP2 knockout zebrafish.

RP2 knockout results in decreased protein levels of GRK1, GNAT1 and GNB1 in zebrafish

Because the major interacting partner of RP2 is ARL3, a small GTPase playing an important role in dissociating the myristoyl-binding protein UNC119 or the prenyl-binding protein PDE6D from their cargos, we considered that RP2 might participate in the intracellular trafficking of myristoylated or prenylated retinal proteins such as the subunits of transducin and phosphodiesterase 6 (PDE6) complex in the phototransduction cascade. GRK1 (rhodopsin kinase) is a well-documented prenylated protein (36) involved in the phosphorylation and inactivation of rhodopsin in the light signal transduction process, and is also a disease-causing gene of autosomal recessive stationary night blindness (37). GNAT1, the alpha subunit of rod transducin, is a myristoylated protein that binds to UNC119. In UNC119-deleted mice, the transport of GNAT1 between the inner segment and the OS was delayed (38). Mutations in GNAT1 have been reported to cause autosomal dominant and recessive stationary night blindness (39,40). The beta subunit of rod transducin, GNB1, has been proven to be an RP2 interacting protein, and its membrane association could be reduced by RP2 RNAi in ARPE19 cells (15,33). For these reasons, we were curious about whether the transport of GRK1, GNAT1 and GNB1 were affected by RP2 ablation in zebrafish.

First of all, we detected the protein levels of GRK1, GNAT1 and GNB1 in our RP2^null zebrafish at the age of 10 days, 2 and 4 months. For GRK1, the protein levels of RP2^null groups decreased to ∼50% of WT controls from 10 dpf to 4 mpf (months post-fertilization; Fig. 5A and B). For GNAT1, in RP2^null zebrafish, the protein levels were reduced from ∼40% (10 dpf) to 30% (4 mpf; Fig. 5A and C), showing an age-related decline pattern. For GNB1, although the protein level of 10 dpf stage was unchanged, significant reductions (∼50%) could be detected at 2 and 4 months of age in RP2^null eyes (Fig. 5A and D). The gamma subunit of rod transducin, GNGT1, is also a prenylated protein (41) like GRK1. But we have not obtained suitable antibody against zebrafish GNGT1 for western bolt yet and could not test it.

Figure 5.

RP2 knockout results in decreased protein levels of GRK1, GNAT1 and GNB1 in zebrafish. (A) Protein levels of GRK1, GNAT1 and GNB1 were detected by western blot in 10 dpf (whole), 2 mpf (eyes) and 4 mpf (eyes) WT and RP2^null zebrafish. Alpha tubulin was used as loading control. Asterisk denotes non-specific bands produced by the anti-GNAT1 antibody. The reductions of GRK1 and GNAT1 could be observed as early as 10 dpf. (B–D) The quantitative data of at least three independent experiments were statistically analysed using two-tailed Student's t-test and shown as mean ± SD. **P < 0.01. The down-regulation of GNAT1 is most remarkable. (E) Quantitative PCR analysis revealed normal mRNA levels of GRK1, GNAT1 and GNB1 in RP2^null zebrafish eyes at 2 months of age. Actb1 was served as endogenous control. The data of three independent experiments were analysed using two-tailed Student's t-test and shown as mean ± SD. **P < 0.01.

In addition, IFT20 (a component of the intraflagellar transport complex) has been reported to be involved in the trafficking of rod and cone opsins from the Golgi body to the base of the cilium (42), and its cellular localization is disrupted in the RNAi and RP2 mutant retinal pigment epithelium cells (12,33). In this study, we detected the protein levels of IFT20 in 4-month-old WT and RP2^null eyes by western blot, and found that unlike GRK1, GNAT1 and GNB1, IFT20 was unaffected by RP2 knockout (Supplementary Material, Fig. S7), which is consistent with the normal localizations of rhodopsin and cone opsins in the RP2^null zebrafish.

To further find out whether the down-regulation of these proteins happened on transcriptional level or post-translational level, we performed quantitative RT-PCR analysis to detect the mRNA levels of these genes in 2-months-old WT and RP2^null eyes. Except for RP2, the mRNA levels of GRK1, GNAT1 and GNB1 were unchanged or even up-regulated (Fig. 5E). These results indicated that post-translational mechanism might be responsible for the reductions of GRK1, GNAT1 and GNB1 caused by RP2 elimination.

Abnormal retinal localizations of GRK1, GNAT1 and GNB1 in the RP2^null zebrafish

After confirming the reductions of protein levels of GRK1 and rod transducin subunits, we examined their retinal localizations in 1- and 5.5-months-old WT and RP2^null retinas by immunofluorescence assay. At the age of 1 month, the fluorescence signal of GNAT1 was markedly reduced in RP2^null retinas (almost background fluorescence left) when compared with WT retinas (Fig. 6A, middle panel). Meanwhile, in RP2^null retinas, the signals of GRK1 and GNB1 were mildly decreased with mis-localizated GRK1 in the outer plexiform layer (Fig. 6A, upper and bottom panels). At the age of 5.5 months, the mislocalization of GRK1 still could be seen (arrows in Fig. 6B, upper panel), and the declines of the GNAT1 and GNB1 were obvious in the RP2 knockout zebrafish (Fig. 6B, middle and bottom panels). Additionally, the retinal localizations of the cone opsins kinase GRK7 were unaffected by RP2 knockout at the age of 2 and 4.5 months (Supplementary Material, Fig. S8), suggesting that the rod degeneration is predominant in our RP2^null zebrafish. Based on the above-mentioned results, we speculated that RP2 might play a role in maintaining the protein level of GRK1 and rod transducin complex, which is important for the function and survival of the rod cells. The abnormal protein levels or retinal localizations of these proteins might be one of the pathogeneses of the rod predominant retinal degeneration caused by RP2 mutations probably.

Figure 6.

The retinal localization of GRK1, GNAT1 and GNB1 in the WT and RP2^null zebrafish at the age of 1 month (dark-adapted; A) and 5.5 months (under background light; B). The arrows in the upper panel indicate the mistrafficked GRK1 protein. Scale bars: 20 µm.

Ablation of RP2 causes aberrant distribution of the farnesylated proteins in the retinas

Recently, mistrafficking of prenylated proteins has been reported to be the cause of RP2 (26). Using an antibody recognizing the farnesyl and geranylgeranyl groups (prenyl groups), we performed immunofluorescence assay on the 2- and 6-months-old retinal sections from WT and RP2^null zebrafish to investigate the distribution of the total farnesylated proteins. We found that in WT retinas, the signals mainly localized to the photoreceptors from the outer plexiform layer to the base of inner segment and were faint in the OS. The fluorescence of the peripheral retina was stronger than that of the middle retina and central retina one by one (Fig. 7A and B, left panels). In 2-month-old RP2^null retina (early stage of retinal degeneration), although no obvious difference could be seen in the peripheral and middle regions, the reduction of the signals near inner segment in the central region was evident (Fig. 7A). Furthermore, in 6-month-old RP2^null retina (middle stage of retinal degeneration), the signals near inner segment were mostly lost in all the three regions (Fig. 7B). These results implied that deletion of RP2 might affect the trafficking of farnesylated proteins in the retina, which might be involved in the progress of the retinal degeneration in our RP2 knockout zebrafish.

Figure 7.

Abnormal retinal distribution of the farnesylated proteins in the RP2 knockout zebrafish. Retinal cryosections from 2 (A) and 6 months (B) WT and RP2^null zebrafish were immunostained with the anti-farnesyl antibody to reveal the localization of the total farnesylated proteins. Central, the regions near the optic disc; peripheral, the peripheral retinas; middle, the regions between the central and the peripheral. The arrows indicate the main positions of the fluorescence signals. Scale bars: 20 µm.

Discussion

Retinitis pigmentosa refers to a group of progressive retinal dysfunction and degeneration diseases with irreversible loss of vision. In the hope of increasing the understanding of RP2-related RP and establishing an animal model for rapid screening of therapies, we generated an RP2 knockout zebrafish line, identified the progression of retinal degeneration in the RP2 deficient model and investigated the function of zebrafish RP2 protein in vivo in this study.

Mutations of human RP2 are thought to cause early-onset night blindness and subsequent peripheral and central visual loss, usually accompanied by high myopia with a semi-dominant X-linked genetic pattern (34,35,43–45). The female carriers have been shown to vary from being asymptomatic to having a tapetal reflex or peripheral pigmentation, or even profound visual loss in later life. Recent clinical data revealed that a lot of RP2 patients show macular involvement in fundus examination and cone dysfunction in ERG recording early in their lives, suggesting a diagnosis of rod-cone dystrophy (8). In the exon2-deleted RP2 mice, significant mislocalization of cone opsins rather than rod opsins occurs at the age of 2 months (25). However, in the gene-trapped RP2 mice, both the cone opsins and rhodopsin target normally to OSs. Instead, mistrafficking of cone PDE6 and GRK1 is more evident than rod PDE6 (26). These results, combined with the phenotype identified in the two RP2 knockout mice, accord with the presentations of rod-cone dystrophy. Our RP2^null zebrafish mentioned above show normal retina development and correct localization of rhodopsin and cone opsins at the age of 1 month. Mild deterioration of the rods' OSs can be noted from as early as 2 months of age, whereas the cones' OSs are not affected. In 4-months-old RP2^null retinas, the OS degeneration of rods becomes apparent, yet the cones are moderately affected. Furthermore, in 7-month-old RP2^null retinas, both rods and cones show aggravated OS degeneration. Our observations prefer the diagnosis of typical RP to rod-cone dystrophy, which is much less severe than the phenotype of RP2 morphants described previously (29,30). The presence of ciliopathies such as pronephric cysts and heart looping defects in RP2 knockdown zebrafish (11) has not been checked yet in our RP2^null zebrafish.

As a GTPase activator protein for ARL3, RP2 is speculated to play an important role in the correct localization of myristoylated or prenylated retinal proteins through the ARL3-UNC119 or ARL3-PDE6D complex for several years (46–49). In the RP2 knockout zebrafish, we found that the myristoylated protein GNAT1 and the prenylated protein GRK1 are both down-regulated, as well as the RP2 interaction protein GNB1. These results partly agree with the observations from the gene-trapped RP2 mouse model, which shows declined level of GRK1 in immunofluorescence assay but with normal protein levels of GRK1 and GNAT1 in western blot assay (26). The three models both show normal transport of rhodopsin, although the OSs of rods are dramatically shortened in late stages. The gene-trapped RP2 mice exhibit a slow progression of retinal degeneration with a ∼15% reduction in the retina thickness at 2 years of age. These may explain the normal levels of GRK1 and GNAT1 in 1-month-old RP2-trapped mouse retinas. Furthermore, there is only one opsin kinase (GRK1) in mouse while human and zebrafish have two opsin kinases, GRK1 and GRK7 (specific for the phosphorylation of cone opsins) (50). The retinal localization of GRK7 is not changed in 2- and 4.5-months-old RP2^null zebrafish eyes (Supplementary Material, Fig. S8). This difference may partly explain the rod-dominant phenotype of RP2 patients and RP2 knockout zebrafish and the cone-dominant phenotype of RP2 knockout mice. Besides, we also found abnormal distribution of the prenylated proteins in RP2 mutant retinas, suggesting that RP2 may play a role in the trafficking of prenylated retinal proteins. The exact mechanism of how RP2 regulates GRK1, GNAT1 and GNB1, whether RP2 interacts with them and if ARL3 is involved in requires further studies in animal and cell line models. According to the best of our knowledge, we guessed that the post-translational modification (such as prenylation and myristoylation) and intracellular trafficking of these proteins may be the key to answer these questions.

Our work provides important cues for revealing the function of RP2 in vivo and the pathogenesis of RP caused by RP2 mutations. In addition, zebrafish is a low-cost laboratory animal model with high rate of spawning and fast development speed. Coupled with the facts that the zebrafish genome has been sequenced completely and many of the genes and pathways are conserved between human and zebrafish, we believe that our RP2^null zebrafish, the first RP zebrafish model constructed by TALEN technology, will be a useful platform for large-scale screening of therapies against RP in future.

Materials and Methods

Zebrafish maintenance and breeding

The study was approved by the Ethics Committee of Huazhong University of Science and Technology. Laboratory inbred wild-type AB line of zebrafish was employed. The zebrafish were placed in recirculating water system (pH 6.6–7.4, 26–28.5°C) with a daily cycle of 14 h of light and 10 h of dark. Adult zebrafish were fed three times per day with fresh brine shrimps, whereas baby fish were fed three times daily with live paramecia after 5 dpf, paramecia mixed with brine shrimps after 10 days and only the brine shrimps after 30 dpf. For mating, males and females (1:1 or 1:2) were moved to crossing cages in the early evening and left undisturbed until the following morning. The males and females in the crossing cage were separated with a plastic divider that was removed in the morning before spawning. The eggs were collected and transferred to a Petri dish in embryo medium, and then kept in an incubator (∼28.5°C) for 72 h until the larvae were hatched.

TALEN construction and microinjection

The online tool TAL Effector-Nucleotide Targeter (TALE-NT) (51) was used to design pairs of TAL effectors (TALE) for the fused endonucleases to target the zebrafish RP2. The targeted sequences are 5′ GTGTGAACTGCCGCATT 3′ (TALE-L) and 5′ CTGAAGAACACACTGCCT 3′ (TALE-R). The assembly of customized TALEs was accomplished using the FastTALE™ TALEN kit (Sidansai Biotechnology, China) according to the operating manual. TALEN mRNAs were in vitro transcribed and purified using mMESSAGE mMACHINE® kit (Ambion). A pair of TALEN mRNAs was mixed and microinjected into one-cell stage eggs of zebrafish.

Mutant zebrafish screening

Ten embryos were collected from each dish of injected eggs at 48 hpf in order to extract a mixed genomic DNA. The effect of the TALEN-mediated mutagenesis was tested by restriction enzyme (AvaII, see Fig. 1A) digestion analysis. The rest embryos of the effective groups were raised to sexual maturity and cross-fertilized with WT fish to balance the genetic background and generate F1 zebrafish. F1 zebrafish were brought up to 2 months and examined using the PCR-restriction fragment length polymorphism (PCR-RFLP) method. The heterozygotes were further subjected to DNA sequencing to identify the exact mutations. Males and females carrying the same mutation mated with each other to make homozygotes. Homozygotes were identified using the PCR-RFLP method and further confirmed by DNA sequencing. Their offsprings were employed in the further research.

Quantitative PCR

Total RNAs were extracted from 2 mpf WT and RP2^null zebrafish eyes using RNAiso Plus reagent (Takara), and reverse-transcribed into cDNA using MMLV reverse transcriptase (Invitrogen) and oligo(dT) primer (Takara). The cDNA was served as templates to detect the mRNA levels of the specified genes by quantitative PCR, which were performed on the StepOnePlus™ real-time PCR System (Life Technologies) using AceQ™ qPCR SYBR® Green Master Mix (Vazyme). The data were analysed using the $2^{-\mathrm{\Delta }\mathrm{\Delta }{\mathrm{C}}_{\mathrm{T}}}$ method in the StepOne software (version 2.3). Significance was determined by two-tailed Student's t-test. Primer sequences for RP2 (NM_213446), actb1 (NM_131031), GRK1a (NM_001034181), GRK1b (NM_001017711), GNAT1 (NM_131868), GNB1a (NM_212609) and GNB1b (NM_213481) were presented in Supplementary material, Table S2.

Western blot

For protein extraction, fresh whole zebrafish (10 dpf) or isolated eyeballs (adult fish) were ultrasonicated in RIPA lysis buffer (100–200 µl) with protease inhibitor cocktail. Protein concentration was determined using the BCA Protein Assay Kit (Beyotime, China). Lysates were mixed with loading buffer and boiled for 5 min at 95–100°C, and then stored at −20°C. Proteins were separated on SDS–PAGE and transferred to nitrocellulose membranes. The membranes were blocked for 1 h at room temperature (RT) in 5% skim milk dissolved in TBST buffer (20 mm Tris–HCl, 150 mm NaCl 0.05% Tween 20, pH 7.6), and then incubated in the dilute solution of primary antibodies (Table 1) overnight at 4°C with gentle agitation. The membranes were washed three times in TBST for 5 min each and incubated in HRP-conjugated secondary antibodies (1:20 000; Thermo) for 2 h at RT. After another three rinses (5 min each) in TBST, the membranes were developed using SuperSignal^® ELISA Femto Maximum Sensitivity Substrate (Thermo) and ChemiDoc XRS+ imaging system (Bio-Rad Laboratories). Quantitative analysis of protein bands was performed by the Quantity One 4.62 software. Significance was determined by two-tailed Student's t-test.

Table 1.

List of primary antibodies used in this study

Antibodies	Source	Antigen	Recognize	Dilution
Anti-RP2	Abclonal, A3212	Human RP2	Zebrafish RP2	1:500 for WB
Anti-RP2	Abclonal, Custom	Zebrafish RP2	Zebrafish RP2	1:1000 for WB; 1:100 for IF
Anti-GAPDH	Proteintech, 10494–1-AP	Human GAPDH	Zebrafish GAPDH	1:3000 for WB
Anti-α-tubulin	Millipore, FCMAB322PE	Human α-tubulin	Zebrafish α-tubulin	1:3000 for WB
Anti-zpr3	ZIRC	–	Rod outer segment in zebrafish	1:200 for IF
Anti-zpr1	ZIRC	–	Green/red double cones in zebrafish	1:200 for IF
Anti-rhodopsin (4D2)	Abcam, ab183399	Bovine rhodopsin	Rod outer segment in zebrafish	1:300 for IF
PNA	Molecular Probes, L21409	–	Cone outer segment in zebrafish	50 µg/ml for IF
Anti-rhodopsin (1D4)	Abcam, ab5417	Bovine rhodopsin	Red-cone outer segment in zebrafish	1:200 for IF
Anti-opn1mw	Aviva, ARP59764	Human OPN1MW	Green-cone outer segment in zebrafish	1:300 for IF
Anti-blue opsin	See acknowledgements	Zebrafish opn1sw2	Blue-cone outer segment in zebrafish	1:100 for IF
Anti-UV opsin	See acknowledgements	Zebrafish opn1sw1	UV-cone outer segment in zebrafish	1:200 for IF
Anti-GRK1	Abclonal, A6497	Human GRK1	Zebrafish GRK1	1:500 for WB; 1:50 for IF
Anti-GNAT1	Abclonal, A6496	Human GNAT1	Zebrafish GNAT1	1:400 for WB; 1:25 for IF
Anti-GNB1	Abgent, AP5036a	Human GNB1	Zebrafish GNB1	1:1000 for WB; 1:100 for IF
Anti-GRK7a	See acknowledgements	Zebrafish GRK7a	Zebrafish GRK7a	1:400 for IF
Anti-IFT20	Proteintech, 13615-1-AP	Human IFT20	Zebrafish IFT20	1:1000 for WB
Anti-Farnesyl	Millipore, AB4073	Farnesyl cysteine	Farnesyl and geranylgeranyl motif	1:50 for IF

ZIRC, Zebrafish International Resource Center; WB, western blot; IF, immunofluorescence.

Electroretinography

The device suitable for zebrafish larvae was set up, and ERG was performed according to previous literatures (52). Briefly, larvae at 7 dpf were placed in a 12-well plate in the afternoon and dark-adapted for at least 30 min prior to positioning them on the reference electrode. Each larva was paralysed by adding a drop of Esmeron (0.8 mg/ml in larval medium; MedChem Express) and oriented properly. The recording electrode was positioned in the approximate centre of the cornea, followed by adaptation in complete darkness for more than 2 min. A 1- to 2-s single stimulus with 6000 lux illuminance was used to make a typical ERG trace. Dozens of traces from a larva were collected in nearly 20 min, and the average of the top five b-wave amplitudes was regarded as the larva's b-wave amplitude. The b-wave amplitudes from RP2^null zebrafish were compared with those from WT group using two-tailed Student's t-test.

Histologic analysis

Isolated zebrafish eyeballs were fixed with 4% paraformaldehyde in PBS overnight at 4°C. For cryosectioning, tissues were cryoprotected in 30% sucrose overnight at 4°C and embedded in OCT compound. For paraffin sectioning, tissues were dehydrated in serial gradients of ethanol, substituted with xylene and embedded in paraffin. Embedded tissues were sliced along the vertical meridian of each eyeball (15 µm for cryosections, 5 µm for paraffin sections). Sections containing the whole retina including the optic disc were stained with hematein and eosin. Photographs were taken under microscope using imaging system.

Immunofluorescence

Immunofluorescence was performed on cryosections. Generally, sections were air dried at RT, washed with PBS for 10 min, permeabilized in PDT (PBS containing 1% DMSO and 0.1% Triton X-100) for 10 min and blocked with 10% normal goat serum in PBDT (PDT containing 1% BSA) for 1 h at RT in a humid chamber. Primary antibodies were diluted to the indicated concentration (Table 1) with 2% normal goat serum in PBDT and added onto the slides, which were incubated overnight at 4°C. The slides were washed three times with PBS and incubated with Alexa Fluor 488 or 594-conjugated secondary antibodies (1:500; Molecular Probes^®) for 1 h at 37°C. For Alexa Fluor 488-conjugated PNA staining, after permeabilization, sections were incubated with PNA in PBS for 1 h at 37°C. DAPI was used to label the nucleus at the concentration of 5 µg/ml. The sections were then rinsed three times with PBS, and mounted with a glycerol-based liquid mountant under coverslips. Fluorescence images were acquired using a confocal laser-scanning microscope (FluoView™ FV1000 confocal microscope, Olympus Imaging). The quantitation of the images was performed using the ImageJ software.

Transmission electron microscopy

Adult zebrafish eyes were enucleated and fixed in 2.5% glutaraldehyde in 0.1 M PBS buffer (pH 7.0) overnight at 4°C. After three washes (15 min each) with 0.1 M PBS buffer, the eyes were further fixed in 1% osmium tetroxide in 0.1 m PBS buffer for 2 h at RT. After three washes (15 min each) with 0.1 m PBS buffer, the eyes were dehydrated in 50, 70, 80, 90, 95 and 100% ethanol successively (20 min each), and incubated in acetone for 20 min at RT. The eyes were treated with 50% (1 h), 75% (3 h) and 100% (overnight) epoxy resin (mixed with acetone, v/v), and then heated at 70°C overnight. Embedded eyes were sliced to ultrathin sections (70 nm) using an Reichert-Jung ultramicrotome (Leica). Sections were stained with 3% uranyl acetate and 3% lead citrate for 15 min and visualized with a transmission electron microscope system (HT7700, Hitachi).

Supplementary Material

Supplementary Material is available here.

Funding

This work was supported by the National Natural Science Foundation of China (nos. 31071106, 81270983, 81371064, 31471199 and J1103514), the Research Fund for the Doctoral Program of Higher Education of China (20120142110077), ‘Program of Introducing Talents of Discipline to Universities’ by Ministry of Education of PR China (B08029), and National Key Technology R&D Program in the 12th Five year Plan of China (2012BAI09B00). The work was also supported by the Rosetrees Trust, UK Fight for Sight and National Research Center (X.S.).

Acknowledgements

We are grateful to Dr Yafeng Liu for his great help in the zebrafish ERG measurement, to Professor David R. Hyde (University of Notre Dame) and Professor Fulton Wong (Duke University, School of Medicine) for the kind gift of the zebrafish UV and blue opsin antibodies and to Professor Stephan C.F. Neuhauss (University of Zurich) for the kind gift of the zebrafish GRK7a antibody.

Conflict of Interest statement. None declared.

Author notes

†

The authors wish it to be known that, in their opinion, the first three authors should be regarded as joint First Authors.