Design of Transcription Activator-Like Effector Nucleases (TALENs) Targeting ahi1 in Zebrafish

Transcription activator-like effector nucleases targeting zebrafish ahi1 were designed with TALENT software (available in the public domain at https://talent.cac.cornell.edu/TALENT/). The Golden Gate assembly method²¹ was used to generate the TALEN constructs. We synthesized 5′-capped mRNA encoding the TALENs using the Sp6 mMESSAGE mMACHINE Kit (Ambion; Thermo Fisher Scientific, Waltham, MA, USA) and microinjected 100 pg of mRNA into wild-type zebrafish embryos at the one cell stage.

Genotyping

The ahi1^lri46 mutants were genotyped by PCR using primers targeting ahi1 exon 5 (5′-TGCAGTCAGGCTGAAGTGTC-3′, 5′-TTTCTTCTTCACTTTGGGTTTGA-3′) to generate a 102 base pair (bp) product (wild-type) or a 95 bp product (mutant). PCR products were resolved on a 3% agarose gel.

In Situ Hybridization

In situ hybridization was performed on fixed tissue from 1-cell stage to 5 days after fertilization (dpf) larvae as previously reported.²² A zebrafish ahi1 cDNA clone was purchased from GE Dharmacon (clone ID 9038651; Lafayette, CO, USA), and antisense probes were generated with the mMESSAGE mMACHINE T7 Kit (Ambion; Thermo Fisher Scientific).

Light and Electron Microscopy

Larvae were bisected through the swim bladder. Tails were used for genotyping while heads were prepared for transmission electron microscopy. Enucleation of adult eyes was performed after euthanasia in fish water at 4°C. Briefly, all tissues were fixed for 1 hour at room temperature in 0.08 M cacodylate buffer containing 2% paraformaldehyde and 2% glutaraldehyde. Samples then were washed with cacodylate buffer and postfixed in 1% osmium tetroxide for 1 hour at 4°C. Samples were washed again and then dehydrated in a graded methanol series before embedding them in Embed-812/DER736 (Electron Microscopy Sciences; Hatfield, PA, USA), using acetonitrile as a transition solvent. Tails were used for genomic DNA extraction and genotyping as described above. Adult animals were genotyped by fin clips at 2 months old. Semithin sections were made with a Leica EM UC7 ultramicrotome (Leica Microsystems GmbH, Vienna, Austria), stained with toluidine blue, and imaged with a Zeiss AxioImager.Z2 (Carl Zeiss Microscopy, Thornwood, NY, USA). Ultrathin sections were stained with uranyl acetate and lead citrate following standard procedures, and electron microscopy was performed on a Tecnai G2 Spirit BioTWIN 20-120 kV digital electron microscope (FEI Company, Hillsboro, OR, USA). Micrographs were acquired with a Gatan image filter and an Orius 832 CCD Camera (Gatan, Inc., Pleasanton, CA, USA).

Optokinetic Response (OKR)

Optokinetic response measurements were conducted between 12 and 6 PM using the VisioTracker system (VisioTracker 302060 Series; TSE Systems GmbH, Bad Homburg, Germany). Contrast sensitivity was assessed as described previously.²³ For the spatial frequency response function,²⁴ the contrast was held constant at 70% and we tested stimuli of 0.02, 0.04, 0.06, 0.08, 0.12, and 0.16 cycles/deg by first increasing and then decreasing the frequency. Each spatial frequency stimulus was presented for 3 seconds before reversing direction for another 3 seconds to minimize saccade frequency. All OKR stimuli were presented with a constant angular velocity of 7.5° per second. The genotypes of individual larvae were confirmed following OKR tests.

Immunohistochemistry and Fluorescence Imaging

Samples were fixed at the designated time points. Fixation protocols varied depending on the primary antibodies being used. For zpr1, zpr3, 1D1, peanut agglutinin (PNA), Ift88 and acetylated tubulin, samples were fixed in 4% paraformaldehyde in ×0.8 PBS at 4°C overnight. For Cep290 staining, heads were fixed in 4% paraformaldehyde in ×0.8 PBS at 4°C for a maximum of 2 hours. For Cc2d2a staining, heads were fixed in 2% trichloroacetic acid (TCA) in double distilled water for 2 hours at room temperature. All samples were cryoprotected in 30% sucrose overnight. Cryosections (10 μM) were cut and dried at room temperature overnight. Blocking solution (1% BSA, 5% normal goat serum, 0.2% Triton-X-100, 0.1% Tween-20 in 1× PBS) was applied for 2 hours in a dark, humidified chamber. Primary antibodies were diluted in blocking solution as follows: zpr1 and zpr3 (1:200; Zebrafish International Resource Center, University of Oregon, Eugene, OR, USA), 1D1²⁵ (1:100; gift from James Fadool), acetylated-α-tubulin (1:5000; Sigma 6-11-B1), Ift88 (1:7500),²⁶ Cep290 (1:100; gift from Iain Drummond, Massachusetts General Hospital, Boston, MA, USA), and Cc2d2a²⁷ (1:20; gift from Ruxandra Bachmann-Gagescu, University of Zurich, Zurich, Switzerland). Conjugated secondary antibodies were purchased from Invitrogen Life Technologies (Carlsbad, CA, USA) and used at 1:500 dilutions and 4′,6-diamidino-2-phenylendole (DAPI; 1:1000) was used to label nuclei. For whole mount analysis of the kidney cilia, 36 hours after fertilization (hpf) embryos were fixed in Dent's fixative (80% methanol, 20% dimethyl sulfoxide [DMSO]) overnight at room temperature. The next day, whole mount immunostaining experiments were performed as follows: embryos were rehydrated in a graded series of MeOH:PBST washes for 15 minutes each at room temperature, and incubated in blocking solution (PBS, 1% DMSO, 0.5% Tween 20, 1% BSA, 10% normal goat serum) overnight at 4°C. Embryos were incubated with a mouse anti-acetylated-α-tubulin primary antibody (1:500) diluted in incubation solution (PBS, 1% DMSO, 0.5% Tween 20, 2% normal goat serum) for a minimum of 36 hours at 4°C. Embryos then were washed 4 times with incubation solution for 30 minutes each and incubated with a fluorophore-conjugated secondary antibody (1:1000) diluted in incubation solution overnight at 4°C. Following antibody staining, embryos were washed again with incubation solution, heads were removed for genotyping, and tails were mounted on slides for imaging. Optical sections were obtained with a Zeiss Axio Imager.Z2 fluorescent microscope fitted with the Apotome.2 for structured illumination (Carl Zeiss Microscopy). ImageJ was used on PNA-stained sections to measure cone outer segment length and on 1D1-stained sections to measure rod outer segment area, as well as to create image panels.

Rescue by mRNA Injection

A cDNA clone encoding full-length, wild-type human Ahi1 was purchased from OriGene Technologies (clone SC114030; Rockville, MD, USA). A codon switch encoding the ahi1^R589X mutation was introduced by site-directed mutagenesis (GeneArt Site-Directed Mutagenesis System; Thermo Fisher Scientific). Capped mRNAs were synthesized using the mMessage mMachine Sp6 transcription kit (Thermo Fisher Scientific). Messenger RNA was injected into 1-cell stage embryos with 0.05% phenol red. Embryos were evaluated for phenotypic rescue by immunostaining kidney cilia in whole 36 hpf embryos with acetylated α tubulin antibodies. As kidney cilia are stereotypically longer than other cilia in the developing embryo, we counted all cilia in the distal kidney longer than 3 μm to ensure that the analysis was tissue-specific.

Distal Pronephric Cilia Are Absent in ahi1^lri46 Mutants

As kidney disease is a common feature of JBTS, and because we observed a minor but significant incidence of kidney cysts in mutant animals, we investigated whether renal cilia were affected during early development. To answer this question, we performed whole-mount immunostaining with antibodies for acetylated α-tubulin and Ift88 on 36 hpf embryos (Fig. 2). In wild-type animals, cilia line the entire pronephric duct, making the structure easily discernible in whole mount embryos (Fig. 2A, arrows, arrowheads). In all ahi1^lri46 mutants examined (n = 11), tubulin immunoreactivity was absent in the distal half of the duct (Fig. 2B). At higher magnification, cilia in the proximal duct of ahi1^lri46 mutants appeared similar to those observed in wild-type samples (Figs. 2C, ​C,2D),2D), but were completely absent in the distal half (Figs. 2E, ​E,2F).2F). Interestingly, the fact that kidney cysts are rarely observed in mutants, despite the universal lack of distal cilia, suggests that the pronephros is largely functional.

Open in a separate window

Figure 2

Loss of ahi1 affects other ciliated tissues. (A, B) Anti-acetylated tubulin whole-mount immunostaining (green) of kidney duct in wild-type and ahi1^lri46 mutants at 36 hpf. Anterior is to the left in both images. Cilia line the pronephric duct (small arrowheads) in wild-type animals. The contralateral duct also can be observed (small arrows). Cilia are absent in the middle and distal pronephros in ahi1^lri46 mutants (arrowhead in [B] indicates the proximal boundary where cilia are missing). (C, D) In the proximal region, mutant cilia are not qualitatively different from wild-type controls. (E, F) Cilia in the distal region of the pronephros are absent in mutant animals when compared to controls. Ift88 immunostaining (red) was used to mark the ciliary transition zone. Scale bars: 50 μm (A, B) and 20 μm (C–F).

Outer Segments of ahi1^lri46 Mutants Are Disorganized

We next completed a histologic analysis to examine retinal morphology of 5 dpf larvae. In transverse semithin plastic sections, the ahi1^lri46 mutants exhibited normal retinal lamination (Figs. 3A, ​A,3B)3B) indicating that retinal development does not require ahi1. At higher magnification, the photoreceptor outer segments of ahi1^lri46 mutants were shorter and lacked the ordered, columnar organization observed in wild-type siblings (arrowheads in Figs. 3C, ​C,3D).3D). This phenomenon was more readily observed in sagittal sections through the extreme posterior pole of the eye, where cone outer segments were visible as a regular array of dark circles in wild-type samples (Figs. 3E, ​E,3G,3G, arrowheads). In ahi^lri46 mutant samples, the highly organized array of outer segments was lost, particularly that of the more distal cone types, and instead many of them appeared more ellipsoid, suggesting that they are misshapen or improperly oriented (Figs. 3F, ​F,3H,3H, white arrowheads).

Open in a separate window

Figure 3

Histologic survey of retinas of 5 dpf wild-type and ahi1^lri46 larvae (A–D) Transverse sections of 5 dpf wild-type and ahi1^lri46 eyes. Photoreceptor outer segments (arrowheads) were shorter and less organized in ahi1^lri46 mutants. (E–H) Sagittal sections of wild-type and ahi1^lri46 eyes, taken at the posterior pole just dorsal to the optic nerve (*). Indicated regions (white dashed boxes) in (E, F) are shown at higher magnification in (G, H). Black arrows indicate the larger caliber, more proximal UV cone outer segments. White arrows indicate the smaller caliber, more distal red, green, and blue cone outer segments. R, rod outer segments. Scale bars: 50 μm (A, B), 10 μm (C, D, G, H) and 25 μm (E, F).

Roughly 2.5% of ahi1^lri46 mutants survived to at least 3 months of age and less than 1% survived to 5 months of age. No qualitative difference was observed in rod photoreceptor outer segments in either 3- or 5-month-old animals when compared to wild-type siblings (Figs. 4A–C). In contrast, fewer cone outer segments were observed in 5-month-old mutants (Figs. 4A, ​A,4C,4C, black arrows). The outer nuclear layer (ONL) in the dorsal retina of adult zebrafish is typically 3 to 4 nuclei thick (Fig. 4A). In 3-month-old ahi1^lri46 mutants, the ONL was largely preserved, but by 5 months it was nonuniform, with considerable thinning in some areas (Figs. 4B–C, white arrows). Interestingly, the photoreceptor phenotypes appeared confined to the dorsal retina. Images of photoreceptors in the ventral retina showed largely preserved rod and cone outer segments and a preserved ONL (Figs. 4D–F).

Open in a separate window

Figure 4

Histologic survey of retinas from wild-type and ahi1^lri46 adults. (A–C) Dorsal and (D–F) ventral views of toluidine blue–stained 1-μm thick sections of wild-type, 3- and 5-month-old ahi1^lri46 mutants. In the dorsal retina, fewer outer segments of UV-sensitive cones (black arrows) and thinning in the ONL (white arrows) was observed at 5 months of age. POS, photoreceptor outer segments; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. Scale bar: 50 μm.

Ultrastructural Analysis of Mutant Photoreceptors Reveals Abnormal Outer Segments

Disorganization of the photoreceptor outer segments was more obvious by transmission electron microscopy of 5 dpf larvae. In wild-type samples, cone photoreceptor outer segments extended toward the RPE in a parallel and organized manner, giving a palisade pattern (Figs. 5A, ​A,5C).5C). In contrast, the cone outer segments in ahi1^lri46 mutants were shorter and highly disorganized (Fig. 5B). In several instances, large amounts of vesiculated material were observed at the base of the outer segments (Fig. 5D). Occasionally, outer segments were aligned perpendicular to the plane of section, such that a transverse section produced a cross-section through the axoneme and outer segment (Fig. 5E), and in other cases we found disc membranes wrapped around the periphery of an outer segment (Fig. 5F). Such abnormalities were never observed in sections of wild-type retinas. Despite these abnormalities, we observed that the photoreceptor axonemal architecture adjacent to the transition zone was unaltered in ahi1^lri46 mutants when compared to wild-type controls (Figs. 5G, ​G,5H).5H). These results indicated that proper organization of disc membranes and outer segment morphogenesis requires Ahi1 function, but that Ahi1 is dispensable for photoreceptor ciliogenesis in zebrafish.

Open in a separate window

Figure 5

Ultrastructural analysis of wild-type and ahi1^lri46 photoreceptors. (A, B) Transmission electron microscopy provided ultrastructural views of wild-type and ahi1^lri46 photoreceptors. Mutant outer segments are short, fragmented, and disorganized. (C) Higher magnification of wild-type photoreceptor showing tightly stacked disc membranes and inner segment. (D) Vesiculation at the proximal end of an outer segment in ahi1^lri46 mutant (arrowhead). (E) Improper orientation of disc membranes. In this field, an axoneme (white arrow) and disc membranes from outer segments of three different cells, labeled 1 to 3, are oriented perpendicular to each other. (F) Several layers of discs are abnormally wrapped around the outer segment (white arrow). (G, H) Cross-sections through the ciliary pocket (cp) show the proper 9 + 0 microtubule arrangement of axoneme. Mitochondria (m) are visible in (G). Y-shaped crosslinkers also are present (arrows in [G, H]). Staining artifacts are visible as diffuse, black deposits in some panels. Scale bars: 1 μm (A–F) and 200 nm (G, H).

Progressive Photoreceptor Degeneration in ahi1^lri46 Mutants

Rod outer segment assembly requires rhodopsin transport and defects in rhodopsin trafficking have been reported for Ahi1-null mutant mice.^15,18 Therefore, we hypothesized that the disorganized photoreceptors in ahi1^lri46 mutants would be associated with mislocalized rhodopsin. Using the monoclonal antibody 1D1, we observed that rhodopsin properly localized to the outer segments in ahi1^lri46 mutants at 5 dpf, comparable to wild-type siblings (Figs. 6A, ​A,6B).6B). Rod photoreceptor density and rod outer segment area, as determined by 1D1 staining, also were similar between wild type and ahi1^lri46 mutant animals (Figs. 6G, ​G,6H).6H). Rhodopsin mislocalization was observed in 3-month-old adults and some rhodopsin immunoreactivity was seen within the thinner ONL at 5 months (Figs. 6D, ​D,6F,6F, arrows). The gap between the rod outer segments and the ONL, which typically is occupied by cones, is practically absent in 5-month-old animals (Figs. 6E, ​E,6F).6F). No significant thinning was observed, however, in mutant rod outer segments when compared to wild-type siblings, confirming our previous observations by light microscopy.

Open in a separate window

Figure 6

Immunohistochemical analysis of rod photoreceptors in ahi1^lri46 mutants. (A, B) Rod photoreceptors stained with 1D1 (red, rhodopsin) at 5 dpf. (C–F) Rod photoreceptors stained with zpr3 (red, rhodopsin) at 3- (C, D) and 5- (E, F) month-old animals. Arrows in (D) and (F) highlight rhodopsin mislocalization. (G) Quantification of rod outer segment area at 5 dpf. (H) Quantification of rod density at 5 dpf. Scale bars: 20 μm (A–B) and 50 μm (C–F).

We next examined cone inner segment morphology by immunolabeling with zpr1, a monoclonal antibody that recognizes arrestin-3 on the cell bodies and inner segments of red-green double cones³¹ and with PNA to label the interphotoreceptor matrix surrounding cone outer segments.^32–34 By zpr1 staining, the ahi1^lri46 cone inner segments showed a tightly packed columnar morphology, equivalent to control samples at 5 dpf (Figs.7A, 7B). Peanut agglutinin lectin staining revealed that the mutant cone outer segments were short (4.3 vs. 6.1 μm; P < 0.001) and misshapen at 5 dpf, consistent with our previous observation by light and electron microscopy (Figs. 7C, ​C,7D,7D, ​D,7M).7M). The total number of cone outer segments, however, was preserved (Fig. 7N), which suggested that loss of ahi1 did not result in cell death at 5 dpf. In ahi1^lri46 adults, cone organization and inner segment morphology appeared normal at 3 months of age, suggesting that cones recovered in surviving animals (Figs. 7E, ​E,7F).7F). By 5 months of age, however, the cone inner segments were highly disorganized and appeared missing in some areas (Figs. 7I, ​I,7J).7J). Similarly, fewer cone outer segments were observed and the morphology of remaining outer segments was abnormal in ahi1^lri46 mutants at 5 months (Figs. 7G, ​G,7H,7H, ​H,7K,7K, ​K,7L).7L). Taken together, these data suggested that Ahi1 is required for early cone morphogenesis, yet cones survive and do not begin degenerating until at least 3 months of age.

Open in a separate window

Figure 7

Immunohistochemical analysis of cone photoreceptors in ahi1^lri46 mutants. (A, B, E, F, I, J) Red-green double cones are labeled with zpr1 (red). Cone inner segments are preserved at 5 dpf and 3 months of age, but missing (arrowheads) at 5 months. (C, D, G, H, K, L) PNA staining (red) revealed that mutants had shorter cone outer segments ([D], arrowheads) at 5 dpf and missing at 5 months ([L], arrowheads). (M) Quantification of cone outer segment lengths between wild-type and ahi1^lri46 mutants at 5 dpf. ***P < 0.001. N, quantification of cone density at 5 dpf.

Visual Performance in Mutant Animals Is Not Compromised

As zebrafish visual function is cone-driven at 5 dpf³⁵ and because cone outer segments were disrupted in ahi1^lri46 mutant retinas, we hypothesized that visual performance of ahi1^lri46 mutants would be compromised. We evaluated visual function by assessing the OKR gain. The OKR gain is defined as the ratio between stimulus velocity and eye velocity, and is dependent on angular stimulus velocity, spatial frequency, and contrast of the moving image.²⁴ Reduced gain is indicative of defective visual performance. Using the VisioTracker system,³⁶ we measured OKR gain while varying either contrast or spatial frequency.²⁴ Contrast sensitivity is a general test of visual function, while spatial frequency assays under normal luminance test the ability of cones to discriminate between two stimuli (e.g., bars) separated at decreasing distances. As previously noted,^23,24 wild-type animals show a linear relationship between gain and the logarithm of contrast. Optokinetic response gain in mutant animals was not significantly different from wild-type controls in response to varying either the stimulus contrast or spatial frequency (Fig. 8), and, therefore, we concluded that visual function in these animals is preserved.

Open in a separate window

Figure 8

Visual function is not affected in ahi1^lri46 mutants at 5 dpf. Optokinetic response of wild-type (open circles) versus ahi1^lri46 homozygous mutants (black squares) at 5 dpf as a function of contrast (A) and spatial frequency (B) measured from smooth pursuit eye movements. Contrast sensitivity test n = 21 per genotype. Spatial frequency test n = 14 per genotype. Error bars: SEM.

Loss of Ahi1 Does Not Affect Components of the Connecting Cilium

As Ahi1 localizes to the basal body in cultured cells,¹⁷ we next interrogated whether loss of Ahi1 affected the localization of components of the basal body and axoneme. We first examined basal body localization using the Tg(centrin-GFP) transgenic line, which expresses a centrin-GFP fusion protein.²⁰ Centrin is a calcium-binding protein that localizes to basal bodies.³⁷ Similar to observations in Ahi1-null mice,¹⁸ basal bodies localized normally to the apical pole of the inner segments of wild-type and ahi1^lri46 mutant photoreceptors (Figs. 9A, ​A,9B).9B). We next evaluated the expression of Ift88, a component of the IFT machinery, and acetylated tubulin, to determine if IFT particles localized to the connecting cilium. In wild-type and ahi1^lri46 mutant photoreceptors, Ift88 immunoreactivity was observed as punctate foci at the base of the connecting cilium (Figs. 9C, ​C,99D).

Open in a separate window

Figure 9

ahi1 is not required for ciliogenesis or transition zone protein localization. (A, B) Tg(centrin:GFP) expression (green) localized to the apical inner segments in wild-type and ahi1^lri46 mutants at 3 dpf. Zpr1 (red) was used to stain cone photoreceptors to visualize the apical boundary. (C, D) Five dpf retinas stained with acetylated tubulin (green) and Ift88 (red) showed ciliary localization of the IFT particle in wild-type and mutant animals. (E, H) The transition zone proteins Cep290 and Cc2d2a (red) exhibited punctate staining in wild-type and mutant animals. Scale bar: 20 μm.

The transition zone (TZ) refers to the region at the proximal end of the cilium and functions as a ciliary gate to regulate protein entry. The Cep290 and Cc2d2a proteins colocalize to the transition zone and are two components of a larger MKS complex.^38–40 Mutations in those genes also lead to JBTS in humans.^41,42 Using a rabbit polyclonal antibody against the N-terminus of Cep290, we detected punctate immunoreactivity in the region consistent with the connecting cilium of wild-type and ahi1^lri46 mutants (Figs. 9E, ​E,9F).9F). When immunostaining for Cc2d2a was performed, we observed strong punctate labeling in the connecting cilium of wild-type samples (Fig. 9G). In ahi1^lri46 mutants, Cc2d2a immunoreactivity was less intense than what was observed for wild-type samples, but the localization was normal (Fig. 9H). These results indicated that the absence of Ahi1 does not compromise the localization of these proteins to the transition zone, but may reduce Cc2d2a stability at the transition zone.

The authors thank the members of the Biological Resource Unit and their care of the zebrafish facility at the Cleveland Clinic.

Supported by National Institutes of Health (NIH) Grants F32EY025145 (J.F.), EY017137-S1 minority supplement (E.M.L.), and EY017037 (B.D.P.), and a Doris and Jules Stein Professorship Award from Research to Prevent Blindness (B.D.P.).

Disclosure: E.M. Lessieur, None; J. Fogerty, None; R.J. Gaivin, None; P. Song, None; B.D. Perkins, None