• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Dev Biol. Author manuscript; available in PMC Dec 16, 2009.
Published in final edited form as:
PMCID: PMC2794694
NIHMSID: NIHMS163058

Retinal homeobox 1 is required for retinal neurogenesis and photoreceptor differentiation in embryonic zebrafish

Abstract

Retinal homeobox (Rx/Rax) genes are essential for the organogenesis of the vertebrate eye. These genes are dynamically expressed in a tissue-specific manner during eye development, suggesting pleiotropic roles. We use a temporally-selective gene knockdown approach to identify endogenous functions for the zebrafish rx genes, rx1 and rx2. Depletion of rx1 over the period of eye organogenesis resulted in severely reduced proliferation of retinal progenitors, the loss of expression of the transcription factor pax6, delayed retinal neurogenesis, and extensive retinal cell death. In contrast, depletion of rx2 over the same developmental time resulted in reduced expression of pax6 in the eye anlage, but only modest effects on retinal cell survival. Knockdown of rx1 specifically during photoreceptor development inhibited the expression of multiple photoreceptor-specific genes, while knockdown of rx2 over this time selectively inhibited the expression of a subset of these genes. Our findings support a function for rx2 in regulating pax6 within the optic primordia, a function for rx1 in maintaining the pluripotent, retinal progenitor cell state during retinal development, as well as selective functions for rx1 and rx2 in regulating photoreceptor differentiation.

Keywords: Zebrafish, rx/rax, Eye, Retina, Progenitor, Photoreceptor, Rod, Cone, Transcription factor, Morpholino

Introduction

Organogenesis of the vertebrate eye begins shortly after neurulation, as bilateral evaginations of the neural tube form the optic primordia (Hilfer, 1983; Schmitt and Dowling, 1994). These structures interact with overlying ectoderm and surrounding perioptic mesenchyme to form the structures of the mature eye, including the neural retina. Retinal progenitors are initially distributed within the neural retina as a pseudostratified neuroepithelium as they proliferate. Progenitor cells then generate the diverse cell classes of the mature retina in a stereotyped sequence of cell-specific terminal mitoses (Altshuler et al., 1991; Hu and Easter, 1999). The mature retina has a laminar arrangement, with light-sensitive photoreceptors in an outer nuclear layer (onl), processing neurons such as bipolar cells and amacrine cells in an inner nuclear layer (inl), and the retinal ganglion cells in a third layer (gcl); these cellular layers are separated by synaptic, or plexiform layers (opl and ipl).

The mechanisms through which the vertebrate eye and retina develop are not entirely understood. However, several homeodomain-containing transcription factors are known to play critical roles in the establishment of the optic primordia, the proliferation of retinal progenitors, the regulation of retinal neurogenesis, and the differentiation of postmitotic retinal cells. For example, pax6 expression is essential for the formation of the optic primordia (Walther and Gruss, 1991) and gain-of-function studies suggest that pax6 also promotes progenitor cell proliferation (Zaghloul and Moody, 2007). Pax6 is expressed in postmitotic retinal ganglion cells and amacrine cells (Hitchcock et al., 1996; Ochocinska and Hitchcock, 2007), where it participates as a member of a network of transcription factors that establish and maintain the differentiated ganglion cell state (Mu and Klein, 2004). Six3, Crx, and Chx10/vsx2, are other examples of genes encoding homeodomain-containing transcription factors that play numerous and pleiotropic roles in these processes (Chow and Lang, 2001; Dyer, 2003; Levine and Green, 2004; Passini et al., 1997; Shen and Raymond, 2004; Wargelius et al., 2003).

The retinal homeobox genes (Rx/Rax) also have important functions related to eye development. The Rx gene product contains conserved domains including an N-terminal octapeptide domain, a home-odomain, an Rx domain, and a C-terminal orthopedia-aristaless-rx domain (Mathers et al., 1997; Pan et al., 2006). In humans, mutations in RX result in anophthalmia or microphthalmia (Voronina et al., 2004), a phenotype predicted by the mouse knockout of Rx (Mathers et al., 1997). Loss of rx3 expression in zebrafish and medaka also results in an eyeless phenotype (Loosli et al., 2003; Rojas-Munoz et al., 2005). Gain-of-function studies have shown that overexpression, or ectopic expression of Rx genes can lead to the expansion of the eye field and formation of ectopic retinal tissue in Xenopus and zebrafish models (Andreazzoli et al., 2003; Chuang and Raymond, 2001; Mathers et al., 1997). During eye organogenesis, Rx appears to function as part of a highly conserved network of factors collectively referred to as the eye field transcription factors; this network also includes Pax6, Six3, Optx2, Tlx, Lhx2, and ET (Zuber et al., 2003).

The expression patterns of Rx genes are dynamic and persist beyond the optic primordia stage. For example, mouse and human Rx are each expressed in retinal progenitors and later in specific cell populations of the inl (Voronina et al., 2004; Mathers et al., 1997). Chicken Rax is similarly expressed in retinal progenitors (Chen and Cepko, 2002), and Xrx1 (Xenopus) is expressed in retinal progenitors and later in photoreceptors (Mathers et al., 1997). In the zebrafish, rx3 becomes restricted to the hypothalamus, retinal pigmented epithelium, and cells of the inl, while rx1 and rx2 persist in retinal progenitors and are then expressed in cone and rod photoreceptors (Rojas-Munoz et al., 2005; Chuang et al., 1999; Nelson et al., 2008). Functions for Rx genes during retinal neurogenesis and retinal cell differentiation are therefore likely, but have been difficult to demonstrate due to the confounding effects of the important early role of Rx genes for eye organogenesis. Functions for Xenopus rx1 (Xrx1) in maintaining retinal stem cell identity were suggested through the use of gain-of-function approaches and a dominant-negative form of rx1 (Zaghloul and Moody, 2007), providing the first indication that predicted, later roles may exist.

A function for Rx genes in regulating photoreceptor differentiation is also predicted, as the Rx protein binds to PCE-1 (photoreceptor conserved element-1) elements found in regulatory regions of photoreceptor-specific genes (Kimura et al., 2000). However, a role for Rx in promoting photoreceptor differentiation in vivo has not been shown. Investigators have instead focused on a group of genes phylogenetically related to Rx, the Rx-like genes that are expressed in photoreceptors and regulate their differentiation. The Rx-like genes are distinct from the Rx genes in that they lack the octapeptide domain (Pan et al., 2006). Rx-like genes are present in human and rat (QRX; Wang et al., 2004), chicken (RaxL; Chen and Cepko, 2002), and Xenopus (Rx-l; Pan et al., 2006), and their gene products have been demonstrated to be stronger than Rx at transcriptional activation of photoreceptor genes (Chen and Cepko, 2002; Wang et al., 2004; Pan et al., 2006). Interestingly, the genomes of two major animal models, zebrafish and mouse, do not contain Rx-like genes, raising questions regarding the regulation of photoreceptor differentiation in these models, as well as regarding the roles of the zebrafish and mouse Rx genes.

The zebrafish genome contains three rx genes, rx1, rx2, and rx3 (Mathers et al., 1997). Of these three, rx1 and rx2 are expressed in the optic primordia, retinal progenitors, and in rod and cone photoreceptors (Chuang et al., 1999; Nelson et al., 2008). In addition, rx1 is expressed sporadically within cells of the rod photoreceptor lineage (Nelson et al., 2008). Rod photoreceptors in teleost fish are generated late in embryonic retinal development, and arise from clusters of proliferating cells dispersed throughout the inl; these progenitors are seeded into the onl where they undergo terminal mitotic divisions and differentiate into rods (Johns, 1982; Otteson et al., 2001; Otteson and Hitchcock, 2003; Raymond, 1985; Raymond and Rivlin, 1987). Functions for zebrafish rx1 and rx2 in retinal progenitors, rod and cone photoreceptors, and within the rod lineage, have not been demonstrated. The PCE-1 element has been identified in the promoter regions of the zebrafish genes encoding rod opsin (Kennedy et al., 2001), UV opsin (Luo et al., 2004), and cone transducin (Kennedy et al., 2007), suggesting that rx activity may be involved in the regulation of these photoreceptor-specific genes.

In the present study we use a temporally-selective loss-of-function approach to identify endogenous functions for the retinal homeobox genes, rx1 and rx2, during retinal neurogenesis in the zebrafish. We show that rx2, but not rx1, regulates pax6 expression in the optic primordia; however, rx1 becomes more important for the maintenance of pax6 expression in retinal progenitors during early retinal neurogenesis. Expression of rx1 is also required for the proliferation and survival of retinal progenitor cells. Finally, we demonstrate selective in vivo functions for rx1 and rx2 in regulating expression of photoreceptor-specific genes.

Materials and methods

Animals and maintenance

All experiments involving animals conformed to the principles adopted by the Society for Neuroscience, and were approved by the University of Idaho Animal Care and Use Committee. Zebrafish embryos were maintained at 28.5 °C as described in Westerfield (2000) on a 14:10 hour light/dark cycle. All embryos used in the following studies were derived from a stock originally purchased from Scientific Hatcheries (now Aquatica Habitats). Unless otherwise indicated, embryos were treated with 0.003% phenylthiourea (PTU) at 12 h post-fertilization (hpf) to prevent melanin synthesis (Westerfield, 2000).

Morpholino, capped mRNA, and BrdU injections

25-mer morpholinos (MOs; Gene Tools, LLC Philomath, Oregon) were designed to target either the ATG translation start site or the first GT splice donor binding site for both the rx1 and rx2 mRNA transcripts (Rojas-Munoz et al., 2005). A 5-mispair MO derived from the rx1 ATG targeted MO was used as a control. Each MO sequence was BLAST searched on the NCBI database against the zebrafish genome to ensure that there would not be any non-specific mRNA/morpholino hybridization. The MOs were injected into the yolk of 1–2 cell stage zebrafish embryos or at 44.5 hpf (as indicated in Results) as described previously (Nasevicius and Ekker, 2000; Stenkamp and Frey, 2003; Summerton, 1999). In some cases we injected MOs in a suspension of 0.1% lipofectamine (Gibco BRL, Gaithersburg, MD) into the head, immediately behind the developing eye (as in Stenkamp et al., 2000). Sequences and concentrations of MOs used to target rx1 and rx2 are shown in Supplemental Table 1 (and see Rojas-Munoz et al., 2005).

Capped rx1 mRNAs were generated using the mMESSAGE mMACHINE kit (Ambion, Austin, TX), according to the manufacturer’s instructions. Full-length rx1 cDNA (the gift of Peter Mathers) was used as a template for mRNA synthesis. The capped mRNA was quantified using spectrophotometry, and size was verified by agarose gel electrophoresis. In order to verify the specificity of rx1 MOs, approximately 1 ng of capped rx1 mRNA was coinjected with rx1 splice site-directed MOs at a final concentration of 75 μM, into 1–2 cell-stage embryos. As the splice site-directed MO has a target site on unprocessed RNA, this strategy for rescue carried no risk for direct MO-mRNA interactions that would confound interpretation of the experiment. Phenotypes were assessed at 53 hpf in sectioned embryos.

BrdU incorporation studies were performed as previously described (Hu and Easter, 1999; Nelson et al., 2008). Briefly, 60 hpf zebrafish embryos were immobilized against the inner edge of a Petri dish and injected, into the yolk, with approximately 1.5 nl of 10 mM BrdU solution suspended in sterile saline. Injection volumes were calculated using spherical geometry, and injections were performed using a pressurized microinjection apparatus and pulled glass capillary tubes. Embryos were returned to system water until fixation at 72 hpf.

Tissue preparation

Dechorionated zebrafish embryos were fixed in phosphate-buffered, 4% paraformaldehyde/5% sucrose for 1 h and either stored in 100% methanol as whole mounts or were washed several times in increasing concentrations of phosphate-buffered sucrose and cryoprotected overnight in 20% sucrose. Embryos to be used for sectioning were then embedded and frozen in 1:2 OCT medium/phosphate-buffered 20% sucrose (Sakura Finetek, Torrance CA) (Barthel and Raymond, 1990). Transverse sections of the embedded embryos were collected on Fisher SuperFrost slides at 5 μm using a cryostat.

Immunocytochemistry

Immunocytochemistry was performed as described previously (Stenkamp et al., 2000). Briefly, slides were incubated for 30 min in a blocking solution containing 10% goat serum, followed by an overnight incubation in primary antibody. This was followed by a 30 minute wash in phosphate-buffered saline plus 0.5% Triton X-100 (Sigma; St. Louis, MO) (PBST) and incubation for 1–2 h in secondary antibody (goat-anti-rabbit or donkey-anti-mouse Cy3 conjugated, 1:200; Jackson ImmunoResearch Laboratories, INC, West Grove, PA).

The slides were then washed for 30 min in PBST and mounted in Vectashield (Vector Laboratories, Burlingame, CA) for visualization using Nomarski and epifluorescence optics on a Leica DMR compound microscope. The following primary antibodies and concentrations were used; mouse monoclonal anti-HuC/D (1:200; Invitrogen Molecular probes, Eugene OR) for labeling amacrine and ganglion cells; rabbit polyclonal anti-PKC (1:2000; Santa Cruz Biotechnologies, INC, Santa Cruz, CA) for labeling bipolar cells; mouse monoclonal anti-glutamine synthetase (1:1000; BD Biosciences Pharmingen, San Diego, CA) for labeling Müller glial cells; rabbit polyclonal anti-phospho-histone H3 (1:1000; Upstate Cell Signaling Solutions, Lake Placid, NY) for labeling cells undergoing mitosis; mouse monoclonals zpr1 (1:200), zpr3 (1:200), and zn8 (1:25; all from the Zebrafish International Resource Center, University of Oregon, Eugene, OR) for labeling cones, rods, and ganglion cells, respectively, and rat anti-BrdU (1:100; Accurate Chemicals, Westbury, NY).

For BrdU immunocytochemistry we acid-treated sectioned material in a 1:1 solution of 4N HCl:PBST for 30 min followed by washes in PBST and PBS (Stenkamp et al., 1997). The sections were then blocked with 20% goat serum for 30 min followed by an overnight incubation with the primary anti-BrdU antibody. These preparations were then washed in PBST and incubated with a secondary donkey anti-rat FITC (1:200, Jackson ImmunoResearch Laboratories, INC, West Grove, PA), washed in PBST and mounted for microscopy to visualize BrdU-positive cells.

Histology and cell death assays

Retinal histology was assessed on fixed and sectioned material using the nuclear stain 0.01% 4′-6-diamidino-2-phenylindole (DAPI; Sigma Aldrich, St. Louis, MO) and epifluorescence microscopy. Cell death was assessed on fixed and sectioned tissue using Terminal dUTP Nick-End Labeling (TUNEL) and detected using an in situ detection kit with a peroxidase amplification step (Roche; Indianapolis, IN). In living embryos, cell death was assessed at 25 and 34 hpf using acridine orange (Sigma). Briefly, dechorionated embryos were immersed in 5 μg/ml acridine orange for 10 min, and then rinsed 5 times in phosphate buffered saline (PBS). The embryos were then transferred to a slide and cell death was visualized using an FITC epifluorescence filter on a Leica DMR microscope (Shen and Raymond, 2004).

In situ hybridization and scoring of expression domains

Digoxigenin-labeled (DIG) cRNA probes were prepared from the following full-length cDNAs using components of the Genius Kit (Roche): zebrafish rod opsin, red cone opsin, blue cone opsin, and UV cone opsin (gifts of T. Vihtelic), gnat1 and gnat2 (rod and cone transducin, respectively; gifts of Q. Liu), pax6, (gift of S. Wilson), rx1 and rx2, (gifts of P. Mathers), crx (gift of P. Raymond), and NeuroD (gift of V. Korzh). In situ hybridization methods for cryosections and whole mounts have been described previously (Barthel and Raymond, 1993; Stenkamp and Frey, 2003).

Heads of whole mounted embryos hybridized with pax6 were scored for the absence of labeling, vs. weak or strong labeling, in identified regions of the pax6 expression domain. The eyes of whole mounted embryos hybridized with an opsin or transducin cRNA were scored for the extent of the expression domain as described previously (Raymond et al., 1995; Stenkamp et al., 2000). In brief, the absence of labeled photoreceptors was scored as stage 0, the presence of a few labeled photoreceptors in ventronasal retina was scored as stage 1, with higher stages indicated a larger labeling domain. Stage 6 represents eyes having widespread labeling of photoreceptors. Sections labeled using immunocytochemical markers for photoreceptors were scored as in Stenkamp et al. (2002). Briefly, sections bisecting the lens were scored for the absence of labeling (=‘none’), the presence of fewer than ten labeled photoreceptors (=‘few’) or the presence of more than ten labeled photoreceptors (=’many’).

Microscopy, photography, statistics

Sections were viewed on a Leica DMR microscope under epifluorescence and/or Nomarski optics and were imaged using a Spot camera and associated software. In some cases, images collected using Nomarski optics were combined with epifluorescence images using the Apply Image function in Adobe Photoshop CS (Mountain View, CA). Whole mounts were viewed and photographed using brightfield optics. Statistics were performed in the R statistical environment (R Core Development Team, 2008).

Reverse transcriptase-polymerase chain reaction (RT-PCR)

Total RNA from 30, 53, 60, and 72 hpf embryos was extracted using the RNeasy kit (Qiagen, Valencia, CA) following the manufacturer’s recommended protocol for animal cells. The total RNA was then treated with RNase-free DNase I (Invitrogen, Carlsbad, CA) for 30 min at 37 °C. The DNase I enzyme was killed using 25 mM EDTA incubated for 10 min at 65 °C. 50 ng of total RNA was reverse transcribed using the LongRange 2Step RT-PCR kit (Qiagen) and poly (dT) primers, and 150 ng of the resulting cDNA was amplified using Qiagen’s LongRange PCR protocol following the manufacturer’s recommendations. The primer sequences used in the PCR amplification step were as described in Rojas-Muñoz et al. (2005). For rx1 the sequences were, 5′-ACAAGGACCAGGATTCGTTG-3′ and 5′-CATGGAGCTGGTATGTGGTG-3′ and for rx2 the sequences were, 5′-CAGAACCACCTTCACCACCT-3′ and 5′-GCTCTGTCCAGGACCCATAA-3′. The presence of splice variants as a result of the morpholino treatment was visualized on a 1% agarose gel with GelDoc system software.

Results

Rx1 depletion results in lamination defects and reduced retinal cell differentiation

Retinal lamination in the developing zebrafish occurs between 40 and 72 hpf, at the completion of which there are three distinctive cellular layers (gcl, inl, and onl) and two plexiform layers (ipl and opl) (Schmitt and Dowling, 1999). By 53 hpf, retinal lamination is evident in untreated embryos and in those injected with a 5-mispair control MO (Fig. 1A; n=12). By 72 hpf lamination is essentially complete (except at the circumferential germinal zone, cgz; Fig. 1C; n=11). The retinal phenotypes of rx1- and rx-2-depleted embryos have been reported (Rojas-Munoz et al., 2005) but not fully characterized. In agreement with these previous results, rx1-depleted embryos showed no lamination at 53 hpf (Fig. 1B; n= 16). At 72 hpf, rx1 morphant retinas had developed plexiform layers, but these were highly reduced compared to the controls, and the onl was thin and irregular (Fig. 1D; n= 7). In addition, the retinas of rx1-depleted embryos appeared smaller than those of controls (Figs. 1A–D). In contrast, rx2-depleted embryos had an unremarkable retinal phenotype when evaluated at 48 and 72 hpf, showing only a slight reduction in eye size (Rojas-Munoz et al., 2005).

Fig. 1
Morpholino-mediated depletion of rx1 results in lamination defects and reduced retinal cell differentiation. Embryos were treated at the 1–2 cell stage with a morpholino cocktail containing ATG (translational start site, 100 μM) and GT ...

To determine whether retinal cell differentiation takes place following rx1 depletion, we probed sectioned morphant retinas with several retinal cell markers. Retinal ganglion cells are the first retinal cell type to differentiate, between 32 and 40 hpf in the zebrafish (Schmitt and Dowling, 1999). We tested a cell surface marker for ganglion cells and their axons, zn-8 (Hu and Easter, 1999) on control and morphant retinas at 53 hpf (control n=8; treated n=5) and 72 hpf (control n=11; treated n=14) (Figs. 1E–H). At both times, expression of zn-8 was patchy as compared to the extensive and continuous expression of this surface antigen in control retinas.

The HuC/D antigen is expressed in retinal ganglion cells and in amacrine cells of the inl (Henion et al., 1996). Rx1-depleted embryos showed a reduction in the expression of this antigen, such that it was restricted to central location in the retinal gcl (Figs. 1J, L). In control embryos, HuC/D was expressed throughout the gcl and the proximal portion of the inl at 53 (control n=10; treated n=14) and 72 hpf (control n=10; treated n=17) (Figs. 1I, K).

The zpr1 antibody labels a cell surface epitope on red/green-sensitive double cone photoreceptors in the zebrafish (Larison and Bremiller, 1990). At 53 hpf, control retinas were heavily stained with zpr1 in the onl (Fig. 1M, n=11) and by 72 hpf this staining was confluent (Fig. 1O, n=7). In rx1-depleted embryos, zpr1 expression was either eliminated, or localized to a small ventral patch in the developing retina of 53 hpf embryos (Fig. 1N, n=14). This limited domain of expression persisted up through 72 hpf (Fig. 1P, n=25). Nearly 85% of morphants exhibited a reduction in zpr1 staining and in approximately 80% of this population there was no detectable zpr1 staining.

The zpr3 antibody labels an antigenic region on the rod opsin protein (Schmitt and Dowling, 1996; Shkumatava et al., 2004). In 53 hpf control retinas (Fig. 1Q, n=11), zpr3 staining was evident throughout most of the onl, and by 72 hpf zpr3 staining was extensive, labeling predominantly the apical regions of developing rods (Fig. 1S, n=13). In rx1-depleted embryos, zpr3 expression was either eliminated, or limited to a small ventral patch, at both 53 hpf (Fig. 1R, n=17) and 72 hpf (Fig. 1T, n=19). At 72 hpf, approximately 90% of rx1 morphants exhibited a reduction in zpr3 staining and in ~80% of this population zpr3 staining was not detectable in sectioned material.

Rod bipolar cells express the enzyme protein kinase C (PKC) and are located in the inl (Koulen et al., 1997; Shkumatava et al., 2004; Stenkamp et al., 2000). PKC was weakly expressed in 53 hpf control retinas and was not detectable in rx1 morphants (data not shown). By 72 hpf, PKC was strongly expressed in the inl of control retinas (n=6;Fig. 1U) but was restricted to a central patch of expression in treated retinas (n=7; Fig. 1V).

Glutamine synthetase (GS) is expressed in Müller glial cells, whose cell bodies are located in the inl (Linser et al., 1985). GS was weakly detected in 53 hpf control retinas and was not detected in rx1 morphants at this developmental time (data not shown). GS was expressed strongly in the retinas of 72 hpf control treated embryos (Fig. 1W; n=5) but was not expressed in rx1 morphants (Fig 1X; n=7).

In order to verify the specificity of the rx1 MOs, we performed rescue experiments. 1–2 cell-stage embryos were injected with 5-mispair control MOs, or with rx1 splice site-directed MOs, or with the splice site-directed MOs together with capped, rx1 mRNA. The control embryos displayed normal lamination, and extensive expression of the cone marker zpr1, as assessed in sectioned embryos at 53 hpf (Fig. 2A; n=7). In contrast, all rx1 MO-treated embryos lacked retinal lamination, and almost all lacked zpr1-positive cells (Fig. 2B; n=12). In the rescue experiments, 11 of 13 embryos receiving a combination of rx1 MOs and capped, rx1 mRNA, exhibited well defined inner and outer plexiform layers as well as multiple zpr1-positive cells (Fig. 2C). These findings confirm that the MOs specifically target the rx1 transcript, and that the eye phenotype in MO-treated embryos is the selective result of rx1 depletion.

Fig. 2
Capped rx1 mRNAs rescue the rx1 depletion phenotype. (A–C) Sectioned embryos processed for indirect immunofluorescence with zpr1 (cone marker), and counterstained with DAPI to reveal retinal lamination. Embryos were injected with 5-mispair control ...

These results show that treating embryos with rx1 directed morpholinos early in development results in an inhibition or delay in the differentiation of multiple retinal cell types, and in lamination failure or delay. Together with the greatly reduced size of morphant eyes, these findings suggest that rx1 may have roles in the control of eye morphogenesis and retinal neurogenesis that are distinct from those of rx2. These potential roles were tested next.

Rx2 depletion selectively affects expressions of pax6 in optic primordia

To establish potential roles for rx1 and rx2 during eye morphogenesis, we analyzed the expression of pax6 mRNA in neural keel-stage embryos following depletion of rx1 vs. rx2. The pax6 protein is a member of the paired-class homeodomain transcription factors and is essential for the development of the zebrafish eye (Nornes et al., 1998), and there is considerable evidence for cooperative regulation of eye development with rx genes. For example, pax6 overexpression in Xenopus results in ectopic retinal tissue in which rx gene expression is upregulated (Chow et al., 1999; Mathers et al., 1997; Mathers and Jamrich, 2000). A reciprocal interaction has also been demonstrated, in that the overexpression of zebrafish rx2, and to a lesser extent, rx1, results in ectopic retinal tissue and increased expression of pax6 (Chuang and Raymond, 2001). The rx and pax6 transcription factors therefore may regulate one another during eye development, and their cooperative activity may be important for the maintenance of retinal stem cells (Zaghloul and Moody, 2007).

Zebrafish embryos depleted of rx1 displayed normal patterns of expression of pax6 mRNA in the eye anlage, and within the midline region of the pax6 expression domain, when assessed at 15 hpf (Figs. 3A, B). However, embryos depleted of rx2 showed noticeably reduced pax6 expression within the eye anlage, as well as in the most anterior portion of the midline expression domain (Fig. 3C). We quantified our results by scoring all embryos according to the presence and/or apparent strength of expression in each of these domains (Figs. 3D–F). The reduced expression within eye anlage and the anterior midline region of rx2-depleted embryos as compared to controls, was statistically significant (Kruskal–Wallis 1-way ANOVA; p<0.001), while a comparison between pax6 expression scores in rx1-depleted embryos vs. controls showed no statistical significance (eye anlage p=0.9289, anterior region p=0.5146). These results suggest that rx2, but not rx1, functions upstream of pax6 during eye morphogenesis in the zebrafish. This is surprising given the apparently weak phenotype observed for rx2 morphants as compared to rx1 morphants at 53 hpf (Rojas-Munoz et al., 2005). Therefore we investigated potential roles for these two transcription factors during retinal neurogenesis.

Fig. 3
Rx2 but not rx1 depletion causes a reduction in pax6 expression in optic primordia. (A–C) Embryos injected with MOs at the 1–2 cell stage and hybridized with a pax6 probe at 15 hpf. (A) 5-mispair control treated embryo (n=48) showing the ...

Rx1 depletion selectively affects expression of specific transcription factors associated with retinal neurogenesis

The expression of the basic helix loop helix transcription factor atonal homologue 5 (ath5), marks the initial production of retinal neurons (Masai et al., 2000), and is essential for the development of retinal ganglion cells in the zebrafish (Kay et al., 2001). In rx1 morphants, ath5 expression was not detected at 25 hpf, a time when control retinas show an extensive expression domain (Figs. 4A, B; control n=22; rx1 morphant n=26). In contrast, rx2 morphants (n=13) showed little or no disruption of ath5 expression (Fig. 4C). These results suggest an abnormal onset of neurogenesis in the case of rx1 but not rx2 depletion. Rx1-depleted embryos assessed at 34 hpf, however, showed a relatively normal pattern of ath5 expression (data not shown), indicating that the onset of retinal neurogenesis was delayed rather than prevented.

Fig. 4
Expression of ath5 and pax6 but not rx2 is disrupted in rx1-depleted embryos. (A–C) Embryos fixed at 25 hpf and processed for in situ hybridization with an ath5 probe show strong retinal labeling in control embryos (A), no labeling in rx1-depleted ...

During retinal neurogenesis, rx1 morphants (n=35) showed a considerable reduction in pax6 expression in the retina (Figs 4D, E) when compared to control embryos (n=25). Expression of pax6 in the lens appeared unaltered. The expression of pax6 in rx2 morphants however, was identical to that of control embryos (Fig. 4F; n=14). These results are consistent with a regulatory relationship for rx1 and pax6 specifically in retinal progenitors, but not for rx2 and pax6. Therefore, while rx2 functioned upstream of pax6 in the optic primordia (Fig. 3), rx1 assumes this role during retinal neurogenesis.

To determine whether the effects on pax6 persisted, and may therefore explain the later defects in neurogenesis, we performed in situs for pax6 in retinal sections derived from older embryos. In control retinas processed at 53 hpf, pax6 was expressed in newly-generated cells of the gcl and inl, as well as in progenitor cells of the peripheral retina (Fig. 4G; n=6). Retinas derived from rx1 morphants contained pax6-expressing cells (Fig. 4H; n=5), indicating that any regulatory effect of rx1 on retinal pax6 is transient, or that the rx1-MOs were no longer effective at 53 hpf (see below). However, the pax6 expression domain of morphant retinas was broader (Fig. 4H); a feature likely related to the inhibition or delay in retinal lamination.

Zebrafish rx2, like rx1, is expressed in retinal progenitors and in cone photoreceptors (Chuang et al., 1999), and throughout early development these two genes have nearly identical expression domains. At 34 hpf, in rx1 morphants (n=21) and control treated embryos (n=16), rx2 expression in the developing retina appeared nearly identical (Figs. 4I, J). We also examined rx2 expression at 72 hpf, when expression of rx1 and rx2 is restricted to the developing onl, scattered cells in the inl, and the retinal progenitor population at the periphery (Fig. 4K; n=7). In rx1 morphant retinas, the expression domain of rx2 was broader, but superficially resembled that of the control retina, with rx2+ cells at the retinal periphery, and in a region of the outer retina lacking laminar definition (Fig. 4L; n=8). These data suggest that rx1 does not regulate expression of rx2, and that any effects of rx1-MO treatment are the results of the selective loss of rx1 activity.

Rx1 depletion alters the time-course of retinal proliferation

We next tested whether rx1-MO or rx2-MO treatment affected retinal cell proliferation during retinal neurogenesis, using antibodies directed against phosphorylated histone-3, a marker for M-phase (Wei et al., 1999). In control retinas examined early in retinal neurogenesis (25 and 34 hpf), mitotic cells were abundant, and nearly exclusively localized to the outer retina (Fig. 5A, C; n=36, n=32). In rx1-depleted retinas, there were significantly fewer mitotic cells (Figs. 5B, D, I; n=33, n=35; p<0.0001). We also observed several instances of disorganized clusters of dividing cells in rx1 morphant retinas (arrow in Fig. 5D). These findings are consistent with a role for rx1 in maintaining a proliferative, stem cell identity. These results may also explain the small size of embryonic eyes in rx1 morphants (Fig. 1). In contrast, rx2 morphants (n=11, n=13) showed no significant difference (p=0.806684, p=0.809966) in retinal cell proliferation at either 25 or 34 hpf (Fig. 5I and data not shown) compared to controls (n=14, n=17). These results suggest that, unlike rx1, rx2 does not play a major role in retinal cell proliferation during early retinogenesis.

Fig. 5
Rx1 morphants but not rx2 morphants exhibit significant changes in retinal progenitor proliferation. (A–H) Retinal cryosections stained with anti-PH3, a marker for cells in M-phase, from embryos fixed at 25 hpf (A, B), 34 hpf (C, D), 53 hpf (E, ...

We next examined cell proliferation in retinal tissues fixed at 53 hpf and 72 hpf, corresponding to late retinal neurogenesis, and the completion of embryonic retinal neurogenesis, respectively (Hu and Easter, 1999; Raymond et al., 1995). At both 53 hpf (n=16 retinas control and treated) and 72 hpf (n=15 retinas control and treated) rx1 morphant retinas had significantly higher numbers of mitotic cells when compared to control retinas (Figs. 5E–H, I; p<0.0001 in each case). These results show that rx1 depletion has dynamic effects on the proliferation of retinal progenitor cells.

Rx1 depletion is lethal to retinal progenitors

The role of rx1 expression for survival of retinal progenitors was evaluated through the use of two cell death assays. Whole embryos were examined at 25 and 34 hpf using acridine orange incorporation, and sectioned embryonic eyes fixed at 53 and 72 hpf were examined using the TUNEL assay. At 25 hpf, the average number of dying cells in rx1 morphant eyes was not statistically significant compared to control treated embryos (Figs. 6A, B, E; n=15, 13, control and treated, respectively, p=0.371865). In contrast, at 34 hpf, the average number of dying cells in rx1 morphant eyes was significantly higher than the number of dying cells in control eyes (Figs. 6C–E; n=13, 12, control and treated, respectively; p<0.0001). Cell death was eye-specific, suggesting that cell death was associated with the lack of rx1 expression in retinal progenitors and not a general toxic effect of the MOs. The average number of dying cells in the eyes of rx2 morphants was not statistically significant at 25 hpf (n=16, n=20, control and treated, p=0.281712) but was at 34 hpf (n=14, n=11, control and treated, p=0.009850), albeit at a lower significance than for rx1 morphants at this time. These findings suggest that rx2 may not be as important for the survival of retinal progenitors and may explain why there are more severe histological abnormalities and smaller eyes in rx1 morphants than in rx2 morphants (Fig. 6E, and data not shown). The retinas of older rx1 morphants also displayed higher levels of cell death as compared to controls (Figs. 6F–J; 53 hpf n=16, 14 retinas, control and treated, respectively; 72 hpf n=10, 12 retinas, control and treated, respectively; p<0.0001 in both cases). Cell death did not appear to be regionally restricted, although TUNEL+ cells were curiously absent from outer retina, where mitotic nuclei reside (see Fig. 5). Cell death in the rx1 morphant retinas is therefore inversely associated with the M-phase of the cell cycle. In addition, several cells in the lens epithelia of rx1 morphants were also apoptotic at each time point that we analyzed (see arrows in Figs. 6G, I; these labeled cells were not included in the counts or statistical analysis). Our findings are consistent with a requirement for rx1 expression in retinal progenitors throughout retinal neurogenesis.

Fig. 6
Rx1 and rx2 morphants exhibit significant increases in retinal cell death. (A–D) Grayscale images of live control (A, C) and rx1 morphant (B, D) embryos stained with acridine orange showing very little cell death at 25 hpf (A, B) but a dramatic ...

Because we observed a dramatic increase in cell death in the eyes of rx1 morphants we wanted to be sure that the MOs were not generally toxic to developing embryos. The concentrations of MOs that we used for these experiments (Supplemental Table 1) have been used previously (Rojas-Munoz et al., 2005). Nevertheless we tested three independent clutches for MO toxicity at 3 days post-fertilization (dpf). We found that the MO treatment (n=130) did not increase embryo mortality compared to 5-mispair-injected (n=118) and uninjected controls (n=75). There was no significant difference in embryo mortality following MO treatment when compared to either 5-mispair injected controls and uninjected controls (p=0.2362; Wilcoxon Rank Sum test with a Bonferroni post-hoc comparison). In addition, morphant embryos allowed to develop to 5.5 dpf displayed normal retinal lamination and distribution of differentiated retinal cell types, although eyes remained quite small (data not shown; n=6–10 for each marker and condition), indicating that retinal development can “recover” once the MOs are inactive (see below). This recovery may be related to the increase in proliferation of retinal progenitors observed at 53 and 72 hpf. This “rebound” phenomenon has been reported previously (Van Raay et al., 2005). Morpholino treatments did not cause appreciable cell death in structures outside of the eye, and doses lower than 10 μM (GT-ATG) did not result in the observed microphthalmic phenotype of the rx1 morphants shown above.

Rx1 morpholino perdurance is limited, but can be effectively administered at late developmental times

Effectiveness and perdurance of rx1-MOs and rx2-MOs were evaluated by qualitative detection of aberrant splicing of the mRNA target following use of the GT-MOs (splice site-blocking MOs; Supplemental Table 1), as in Rojas-Muñoz et al. (2005) (Morcos, 2007). RT-PCR with intron-flanking primers amplified products of 276 or 313 bp (indicating the presence of spliced rx1 and rx2 transcripts, respectively) from untreated embryos, or from embryos injected at the 1–2-cell stage with 5-mispair control MOs (Fig. 7). Based upon sequence alignment we predicted unspliced variants to be approximately 4.8 kb for rx1 and 3.7 kb for rx2. In embryos injected with GT-MOs at the 1–2-cell stage, amplicons of these predicted sizes were detected at both 30 hpf and 53 hpf (Fig. 7). However, some normally-spiced transcript was present at 30 hpf, and appeared to increase relative to the aberrantly-spliced transcript at 53 hpf, suggesting incomplete functional knockdown of the target genes at these times, particularly at 53 hpf. In addition, only the normally-spliced transcripts were present at 60 hpf and 72 hpf (Fig. 7 and data not shown), indicating that target gene knockdown was transient. Therefore, the morphant phenotypes described so far are the consequence of depleted rx1 or rx2 predominantly during eye morphogenesis and early retinal neurogenesis.

Fig. 7
Efficacy and perdurance of rx1 and rx2 splice site-targeting MOs in embryos treated at the 1–2-cell stage (“0 hpf”) and at 44.5 hpf. The presence of large amplicons (upper row) indicates aberrant splicing. Lanes 1 and 2, embryos ...

Because we were interested in understanding any roles for rx genes specifically for photoreceptor development, we performed MO injections at an alternative time: 44.5 hpf. We have previously shown that use of MO-based knockdown strategies can be effective when the MOs are delivered at late developmental times (e.g. 50 hpf for shh-MOs;Stenkamp and Frey, 2003), and that MOs gain access to developing cells (Stenkamp and Frey, 2003). In the present study, rx gene expression was also influenced by late MO treatment, as aberrantly-spliced transcripts of 4.8 kb (rx1) and 3.7 kb (rx2) were consistently amplified from embryos injected with rx1-MOs or rx2-MOs at 44.5 hpf (Fig. 7) and evaluated at 53 hpf. Indeed, the ratios of unspliced to spliced transcript at this time appeared higher following late MO injection than following injection at the 1–2-cell stage (Fig. 7, compare lanes 5 and 7), and unspliced transcript occasionally persisted until 60 hpf (data not shown). However, normal rx gene splicing was again evident by 72 hpf (data not shown), suggesting that the late MO treatments cause highly transient effects. We performed some experiments using delivery of MOs suspended in lipofectamine, to the head mesenchyme behind the eye (Stenkamp et al., 2000), and found similar results in terms of MO effectiveness and perdurance, as well as phenotype. These data demonstrate that injection of rx MOs at late developmental times selectively reduces expression of rx genes over the period of photoreceptor neurogenesis and early differentiation.

Temporally-selective depletion of rx1 vs. rx2 causes specific effects on photoreceptor development

We made use of this late MO injection strategy for our remaining experiments, in order to probe the function of rx gene expression specifically during the time of photoreceptor neurogenesis and differentiation. 5-mispair control and either rx1 or rx2 directed MO-treated embryos were allowed to survive ~27 h post-injection (until 72 hpf) and were then fixed and processed as whole mounts for in situ hybridization. In these ‘late rx1 morphant’ embryos, rod and cone opsins and transducins showed a reduction in expression compared to those treated with the control MO (Fig. 8; n=10–15 for each probe and condition). ‘Late rx2 morphant’ embryos showed a reduction in rod and red opsin as well as a reduction in gnat1 and gnat2 expression (Figs. 9A–D, G, H, K, L; n=7–15 for each probe and condition). However, there was little or no reduction in UV and blue opsin expression in late rx2 morphants (Figs. 9E, F, I, J; n=17–30 for each probe and condition) suggesting a selective role for rx1 in the regulation of short wave length cone opsins. In order to determine if each treatment (either rx1 or rx2 MOs) caused a statistically significant (p<0.05) reduction in opsin or transducin expression compared to the control treatment we scored individual, hybridized eyes using the opsin expression stages of Raymond et al. (1995; see also Prabhudesai et al., 2005; Fig. 10). We then performed a Wilcoxon Rank sum test to determine if we could reject the null hypothesis that the late treatment of rx1 or rx2 directed MOs does not cause a reduction in opsin or transducin expression. We found that the expression of all opsins and transducins in the late rx1 morphants were significantly different from that of the control treated group (red, blue, UV, rod, and gnat1 p<0.0001; gnat2 p<0.00173). In the late rx2 morphants, the difference in expression of red and rod opsin and gnat1 and gnat2 were statistically significant (p<0.001). However, for blue and UV opsin we failed to reject the null hypothesis (blue opsin p=0.6772; UV opsin p=0.761). We applied a Kruskal–Wallis ANOVA with a Bonferroni multiple comparison to determine which treatment, rx1 or rx2 directed morpholinos, caused a greater reduction in opsin or transducin expression in late morphant embryos. Our results indicated that late rx1 morphants showed a greater reduction in blue and UV opsin expression that was statistically significant. All of the other comparisons yielded no significant difference.

Fig. 8
Temporally-selective rx1 depletion results in a reduction in the expression of rod and cone opsins and rod and cone transducins. Embryos were injected with the 5-mispair control MO (A, C, E, G, I, K) or an rx1-targeting MO cocktail (B, D, F, H, J, L) ...
Fig. 9
Temporally-selective rx2 depletion results in a selective reduction in the expression of rod and red cone opsins and rod and cone transducins. Embryos were injected with the 5-mispair control MO (A, C, E, G, I, K) or an rx2-targeting MO cocktail (B, D, ...
Fig. 10
Histograms representing the size of the expression domains of photoreceptor-specific genes following temporally-selective rx1 or rx2-depletion confirm selective effects of the retinal homeobox genes. Gene expression stages (=photoreceptor recruitment ...

Dual knockdown of rx1 and rx2 more robustly attenuates rod and cone development but does not inhibit expression of NeuroD or crx

Using MOs directed at either rx1 and rx2 transcripts, or both, we performed late (44.5 hpf) injections and then examined retinal histology. In each case, DAPI staining revealed the presence of an onl containing presumptive photoreceptors, and the retina in general appeared histologically normal (Figs. 11A, B, and data not shown). We next evaluated rod and cone photoreceptor differentiation using zpr3 and zpr1 antibodies on sectioned material. With all experimental treatments we observed a reduction in both zpr3 and zpr1 staining (Figs. 11C–F). We quantified these results using the scoring method of Stenkamp et al. (2002), in which sections bisecting the lens are scored for the absence or presence of few (<10) or many (>10) labeled photoreceptors (see Materials and methods). The greatest frequency of reduction in rod and cone differentiation was observed following dual knockdown of rx1 and rx2 (14 of 21, Fig. 11M), as compared with depletion of rx1 only (8 of 17, Fig. 11K) and rx2 only (13 of 34, Fig. 11L) over the time of photoreceptor development. To determine whether late rx gene knockdown affected the differentiation of other cell types, we labeled control and treated retinas with markers specific for retinal ganglion cells (anti-zn8; n=8ea), ganglion cells and amacrine cells (anti-HU C/D; n=5ea), rod bipolar cells (anti-PKC n=10ea), and Müller glia (anti-glutamine synthetase; n=12ea). We did not observe any variation in staining following late rx1/rx2 MO treatment with any of the cell specific markers indicated (data not shown).

Fig. 11
Embryos subjected to temporally-selective (at 44.5 hpf) depletion of rx1 and rx2 show a significant reduction in photoreceptor differentiation. (A–B) DAPI-stained retinal cryosections obtained from control (A) and rx1/rx2-depleted (B) embryos ...

The zebrafish rx1 and rx2 genes are expressed in developing rod and cone photoreceptors (Chuang et al., 1999; Nelson et al., 2008) and cells of the embryonic rod photoreceptor lineage (Nelson et al., 2008). Each of these cell populations also expresses the photoreceptor transcription factors crx and NeuroD (Hitchcock et al., 2004; Ochocinska and Hitchcock, 2007; Shen and Raymond, 2004), and cooperative activity of Rx and Crx has been demonstrated to regulate photoreceptor-specific genes in cell-free systems (Furukawa et al., 1997; Wang et al., 2004). We examined crx or NeuroD expression in late rx morphant embryos using in situ hybridization on sectioned retinas. In material obtained from six treated retinas that showed a reduction in both zpr3 and zpr1 staining, we observed no disruption of either crx or NeuroD expression in late morphants as compared to controls (Figs. 11G–J). This finding suggests that Rx activity does not regulate expression of either crx or NeuroD during photoreceptor development.

The dual knockdown of rx1 and rx2 does not affect proliferation of rod progenitors

The rod photoreceptors of teleost fish are generated from a population of slowly dividing progenitor cells sequestered in the inl, which migrate to the onl for a final round of cell division prior to differentiating as rod photoreceptors (Johns, 1982; Julian et al., 1998; Otteson et al., 2001). We have recently described the sequence of gene expression within the rod lineage of embryonic zebrafish, and observed rx1 expression within this lineage (Nelson et al., 2008). In order to determine if rx1 and/or rx2 play roles in the proliferation of rod progenitors we treated embryos with rx1 and rx2 MOs at 44.5 hpf, followed by BrdU incorporation at 60 hpf (using the ‘cumulative labeling scheme’ of Nelson et al., 2008; see Materials and methods). In control and rx1/rx2-MO-treated retinas, BrdU-positive cells were observed within the retinal margin, and in scattered cells in the inl and onl (Fig. 12), consistent with our previous report (Nelson et al., 2008). We counted the numbers of BrdU-positive cells within the inl and onl, excluding the region within six cell diameters of the retinal margin, as this region of the onl contains cone progenitors as well as rod progenitors (Nelson et al., 2008; see also Ochocinska and Hitchcock, 2007; see dotted lines in Figs. 12A, B). The numbers of BrdU-positive cells in the inl or the onl of rx1/rx2 morphants (n=11) were not significantly different (one-way ANOVA) from those of control embryos (n=13; Figs. 12A–C). These results suggest that neither rx1 nor rx2 regulate the proliferation of rod progenitors. Instead, rx1 and rx2 may play roles in the differentiation of these progenitors and/or precursors into mature photoreceptors.

Fig. 12
Cell proliferation related to the rod lineage is unaltered following temporally-selective depletion of rx1 and rx2. (A–B) Control (A) and rx-depleted (at 44.5 hpf; B) embryos were treated with BrdU at 60 hpf, and fixed at 72 hpf for BrdU indirect ...

Discussion

The principal findings of this study can be summarized as follows. 1) Rx2 selectively regulates expression of pax6 within the optic primordia. 2) Expression of the rx1 gene, but not that of rx2, is needed for the regulation of retinal progenitor cell proliferation. 3) During retinal neurogenesis, rx1 regulates the retinal expression of other transcription factors required for neurogenesis, pax6 and ath5. 4) Expression of the rx1 and rx2 genes is needed to promote the differentiation of rod and cone photoreceptors. 5) The rx1 and rx2 genes do not regulate expression of other transcription factors required for photoreceptor differentiation, crx and NeuroD. 6) The gene products of rx1 and rx2 each have distinct, but dynamic and pleiotropic in vivo functions. 7) Antisense morpholino oligonucleotides can be used to achieve temporally-selective knockdown of expression of target genes.

Rx2 functions upstream of pax6 during eye morphogenesis, but rx1 is required during retinal neurogenesis

The present work significantly expands on previous findings suggesting that rx gene expression establishes the early eye field (Chuang and Raymond, 2001). Overexpression of rx2, and to a lesser extent rx1, causes an expansion of retinal tissue at the cost of forebrain tissue. Given the profound consequences of overexpression, the general phenotype of the rx1 or rx2 morphant was surprisingly mild (Rojas-Munoz et al., 2005; Fig. 1). The reduced expression of rx1 results in small eyes and lamination defects, while reduced expression of rx2 resulted in slightly smaller eyes but no histological errors (Rojas-Munoz et al., 2005). In the present study we have further characterized these phenotypes and established their mechanistic bases. In the rx1-depleted embryos, not only is lamination disrupted, but the differentiation of multiple retinal cell types is impeded or delayed, consistent with a role for rx1 in retinal neurogenesis. In rx2-depleted embryos, pax6 expression is impaired during the formation of optic primordia.

The rx2 gene product regulates pax6 expression in the eye anlage and in the most anterior region of the midline pax6 expression domain, placing rx2 upstream of pax6 during eye morphogenesis. This is consistent with the observation that overexpression of rx2 leads to a higher incidence of ectopic retinal tissue than the corresponding rx1 gain-of-function (Chuang and Raymond, 2001). In mammals, Rx also functions upstream of Pax6, in that the upregulation of Pax6 (though not its initiation) requires Rx activity (Zhang et al., 2000). It is interesting that later aspects of eye development proceeded relatively normally in the rx2-depleted embryos. The rx2 morphants displayed normal patterns of expression of pax6 during retinal neurogenesis, and only mild defects in lamination (Rojas-Munoz et al., 2005). It is possible that, in these morphants, the activity of rx1, or of one or more alternative eye field transcription factors, is able to compensate for the reduction in expression of both rx2 and pax6 in the optic primordia. Dual knockdown experiments (rx1-MO+rx2-MO) suggest that the latter is likely the case, as the eye phenotype of rx1 and rx2-depleted embryos matches that of rx1-depleted embryos (Rojas-Munoz et al., 2005, and data not shown). Therefore, in the zebrafish, while rx1, rx2, and pax6 are all sufficient to promote tissue fates related to eye morphogenesis, they are apparently not required. Only rx3 has been demonstrated to have this function (Loosli et al., 2003).

The role for rx1 in retinal neurogenesis appears to be several-fold: rx1 regulates the proliferation of retinal progenitor cells, promotes their survival, and regulates the expression of other transcription factors that are required for neurogenesis. Rx1 morphants displayed drastically reduced proliferation of progenitor cells during early retinal neurogenesis, and this effect preceded a massive wave of retinal cell death. The rx1 morphants also failed to express pax6 specifically within the retina, and did not initiate expression of ath5 at the normal time of onset of retinal neurogenesis (Masai et al., 2000). Collectively these data suggest that rx1 may be needed to maintain retinal progenitor cell identity, in part through the regulation of pax6. Furthermore, if this identity is not maintained, retinal progenitors may withdraw from the cell cycle prematurely and engage a cell death program.

Functions for rx genes in regulating the cell cycle and consequently the numbers of retinal progenitors are inferred from gain-of-function studies in Xenopus (Casarosa et al., 2003). In addition, the use of a dominant-negative rx1 in the same animal model results in reduced proliferation within the optic cup, and reduced size of the optic cup itself (Zaghloul and Moody, 2007). Our results demonstrate that rx1 continues to be required for retinal progenitor proliferation during early retinal neurogenesis. Interestingly, in the rx1-depleted zebrafish retinas, reduced proliferative activity was followed by apoptosis. To our knowledge, this is the first report of a role for rx genes in promoting the survival of retinal progenitors.

A cooperative relationship between rx genes and pax6 is important for the establishment of the optic primordia in zebrafish (Chuang and Raymond, 2001; and the present study). However, a temporally-selective approach to manipulation of rx1 gene expression in Xenopus was unable to confirm that regulatory interactions between pax6 and rx1 exist beyond the neural plate stage (Zaghloul and Moody, 2007). It is possible that rx1 must be continuously depleted over developmental time in order to influence pax6 expression in retinal progenitor cells. Alternatively, a co-regulatory relationship between pax6 and rx genes may exist in both Xenopus and zebrafish, but the Xenopus and zebrafish rx genes may show differential functional divergence. Our results indicate that, in zebrafish, a regulatory relationship between rx1 and pax6 persists in retinal progenitors during early retinal neurogenesis. In addition, our data show that normal expression of rx1 is important for ath5 expression and the onset of retinal neurogenesis.

Rx1 is required for rod and cone photoreceptor differentiation

We observed a significant decrease in the expression of opsin and transducin genes following temporally-selective knockdown of rx1 or rx2 during photoreceptor development. This effect is consistent with in vitro studies demonstrating that Rx binding sites are present on the promoter of several opsin genes in Xenopus and zebrafish (Batni et al., 1996; Kennedy et al., 2007; Kennedy et al., 2001; Mani et al., 2001; Moritz et al., 2002) and suggests a role for these factors in the transactivation of opsin promoters in vivo. The zebrafish genome contains three rx genes with dynamic and overlapping but distinctive spatiotemporal expression patterns within the developing eye (Chuang et al., 1999). Rx1 and rx2 in particular are expressed in embryonic cone photoreceptors (Chuang et al., 1999; Nelson et al., 2008), embryonic rod photoreceptors, and cells of the rod photoreceptor lineage (Nelson et al., 2008). In contrast, mammalian genomes contain only one Rx gene (Mathers et al., 1997), and this gene does not appear to be expressed within photoreceptors. However, Rx-like genes have been identified in several vertebrates including human (QRX; Wang et al., 2004), chick (cRaxL; Ohuchi et al., 1999), and Xenopus (Pan et al., 2006). These Rx-like transcription factors are able to function as strong transcriptional activators of photoreceptor specific genes and appear to play a functional role in photoreceptor differentiation (Chen and Cepko, 2002; Pan et al., 2006). A BLAST search of the zebrafish genome against the Xenopus Rx-L amino acid sequence did not yield any hits, suggesting that rx1, rx2, and rx3 are the only retinal homeobox-related genes in the zebrafish. Results of the present study indicate that the rx1 and rx2 genes of the zebrafish may serve functions homologous to the rx-like genes in other vertebrates. The genome of the mouse also lacks an Rx-like gene; in mice, the single Rx gene appears to assume the functions of all of the Rx and Rx-like genes in other vertebrates (Wang et al., 2004).

We recently demonstrated that rx1 is expressed within the proliferative lineage of cells that gives rise to rod photoreceptors within the embryonic zebrafish retina (Nelson et al., 2008). However, rx1 does not appear to be involved in proliferation of rod progenitors, in that temporally-selective rx1 depletion does not affect BrdU incorporation into cells of the lineage. We note that the stem cells residing at the apex of the rod lineage are distinct from other retinal progenitors in that they express pax6 but not rx1, and that committed rod progenitors sporadically express rx1 but not pax6 (Nelson et al., 2008), suggesting that any cooperative, regulatory relationship between these two transcription factors is not maintained within progenitors that are not pluripotent. An alternative mechanism – unrelated to rx genes or pax6 – may therefore be responsible for regulating proliferation of rod progenitors.

The phenotype of morphants following temporally-selective depletion of rx1 and rx2 during photoreceptor development strongly resembles that of embryos lacking the sonic hedgehog gene (syu-/-;Stenkamp et al., 2002). The syu-/- mutants display profoundly reduced photoreceptor differentiation, including reduced expression of multiple opsin genes and both transducin genes (Stenkamp et al., 2002; C. Stevens, K. Russo, D. Stenkamp, unpublished observations), but nevertheless show a well-defined onl that expresses crx and NeuroD (Stenkamp et al., 2002). Interestingly, the syu-/- retinas fail to express the rx1 gene within the onl, although they do express crx, NeuroD, and rx2 (Stenkamp et al., 2002). Together with the results of the present study, these data are consistent with rx1 expression mediating the photoreceptor differentiation-promoting effects of shh signaling.

Subfunctionalization of rx1 and rx2

A genome-wide gene duplication took place during the evolution of teleost fish, and many of the duplicated genes persisted as the duplicates assumed divergent functions (Mathers et al., 1997; Postlethwait et al., 2004; Semon and Wolfe, 2007). The expression patterns of zebrafish rx1 and rx2 are virtually identical (Chuang et al., 1999), but our loss-of-function studies have revealed distinct functions during early and late retinal development. Rx1 is needed for maintenance of retinal progenitor proliferation and the onset of retinal neurogenesis, regulates expression of pax6 during this time, and promotes retinal cell survival. In contrast, the loss of rx2 has a significant influence on expression of pax6 during early eye morphogenesis and has a modest effect on cell survival that together do not generate a remarkable retinal phenotype. The present findings are not the consequence of low efficacy or perdurance of the rx2 MO; if anything, the rx2 MO appears to have greater perdurance than the rx1 MO. These results, together with those of Chuang and Raymond (2001), suggest that rx2 has the greater capacity to promote eye-specific fates, perhaps through early regulation of pax6, while rx1 is more important in maintaining the identity and proliferative activity of retinal progenitors later in development. Rx1 does not appear to regulate expression of rx2, consistent with independent roles for these retinal homeobox genes.

The functions of rx1 and rx2 expressed in photoreceptor cells are also distinct. Temporally-selective depletion of rx1 resulted in reduced expression of multiple opsin genes and both transducin genes, while knockdown of rx2 selectively affected expression of the longer-wavelength opsins and the transducins, with no significant effect for expression of blue or UV opsin. The presence of rx1 mRNA in all photoreceptor types has been demonstrated (Nelson et al., 2008), while the presence of rx2 mRNA has not, and so it is possible that rx2 is simply not expressed in blue and UV cones. Alternatively, the two rx genes may have selective target regions with respect to photoreceptor gene regulation. It is noteworthy that exogenously-applied retinoic acid also has targeted effects on rod and red opsin (upregulation) vs. blue and UV opsin (downregulation) (Prabhudesai et al., 2005), consistent with differential mechanisms for the regulation of opsin genes.

Supplementary Material

Supplemental Information

Acknowledgments

The authors would like to thank members of the zebrafish community for the following cDNAs: the zebrafish opsins (T. Vihtelic); the zebrafish transducins (Q. Liu); pax6 (S. Wilson); NeuroD/nrd (V. Korzh); crx (P. Raymond); and rx1 and rx2 (P. Mathers). We note special thanks to Ruth Frey for providing her technical expertise on the microinjection procedure, and to Baekcheol Choi and Christopher Williams for their help with the statistical analyses. This work was supported by NIH R01EY012146 (DLS), graduate fellowships (SMN) available through NIH P20RR015587 (Idaho INBRE) and the University of Idaho’s BANTech Strategic Initiative, and by an undergraduate research grant (LP) from the Department of Biological Sciences.

Footnotes

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ydbio.2008.12.040.

References

  • Altshuler D, Turner D, Cepko C. Specification of cell types in the vertebrate retina. In: Lam D, Shatz C, editors. Development of the Visual System. MIT press; Cambridge: 1991. pp. 37–58.
  • Andreazzoli M, Gestri G, Cremisi F, Casarosa S, Dawid IB, Barsacchi G. Xrx1 controls proliferation and neurogenesis in Xenopus anterior neural plate. Development (Camb., England) 2003;130(21):5143–5154. [PubMed]
  • Barthel LK, Raymond PA. Improved method for obtaining 3-microns cryosections for immunocytochemistry. J. Histochem. Cytochem. 1990;38(9):1383–1388. [PubMed]
  • Barthel LK, Raymond PA. Subcellular localization of alpha-tubulin and opsin mRNA in the goldfish retina using digoxigenin-labeled cRNA probes detected by alkaline phosphatase and HRP histochemistry. J. Neurosci. Methods. 1993;50(2):145–152. [PubMed]
  • Batni S, Scalzetti L, Moody SA, Knox BE. Characterization of the Xenopus rhodopsin gene. J. Biol. Chem. 1996;271(6):3179–3186. [PubMed]
  • Casarosa S, Amato MA, Andreazzoli M, Gestri G, Barsacchi G, Cremisi F. Xrx1 controls proliferation and multipotency of retinal progenitors. Mol. Cell. Neurosci. 2003;22(1):25–36. [PubMed]
  • Chen CM, Cepko CL. The chicken RaxL gene plays a role in the initiation of photoreceptor differentiation. Development (Camb., England) 2002;129(23):5363–5375. [PubMed]
  • Chow RL, Lang RA. Early eye development in vertebrates. Ann. Rev. Cell Dev. Biol. 2001;17:255–296. [PubMed]
  • Chow RL, Altmann CR, Lang RA, Hemmati-Brivanlou A. Pax6 induces ectopic eyes in a vertebrate. Development (Camb., England) 1999;126(19):4213–4222. [PubMed]
  • Chuang JC, Raymond PA. Zebrafish genes rx1 and rx2 help define the region of forebrain that gives rise to retina. Dev. Biol. 2001;231(1):13–30. [PubMed]
  • Chuang JC, Mathers PH, Raymond PA. Expression of three Rx homeobox genes in embryonic and adult zebrafish. Mech. Dev. 1999;84(1–2):195–198. [PubMed]
  • Dyer MA. Regulation of proliferation, cell fate specification and differentiation by the homeodomain proteins Prox1, Six3, and Chx10 in the developing retina. Cell cycle (Georgetown, Tex) 2003;2(4):350–357. [PubMed]
  • Furukawa T, Morrow EM, Cepko CL. Crx, a novel otx-like homeobox gene, shows photoreceptor-specific expression and regulates photoreceptor differentiation. Cell. 1997;91(4):531–541. [PubMed]
  • Henion PD, Raible DW, Beattie CE, Stoesser KL, Weston JA, Eisen JS. Screen for mutations affecting development of Zebrafish neural crest. Dev. Genet. 1996;18(1):11–17. [PubMed]
  • Hilfer SR. Development of the eye of the chick embryo. Scanning Electron Microsc. 1983;(Pt 3):1353–1369. [PubMed]
  • Hitchcock PF, Macdonald RE, VanDeRyt JT, Wilson SW. Antibodies against Pax6 immunostain amacrine and ganglion cells and neuronal progenitors, but not rod precursors, in the normal and regenerating retina of the goldfish. J. Neurobiol. 1996;29(3):399–413. [PubMed]
  • Hitchcock P, Ochocinska M, Sieh A, Otteson D. Persistent and injury-induced neurogenesis in the vertebrate retina. Progress Retin. Eye Res. 2004;23(2):183–194. [PubMed]
  • Hu M, Easter SS. Retinal neurogenesis: the formation of the initial central patch of postmitotic cells. Dev. Biol. 1999;207(2):309–321. [PubMed]
  • Johns PR. Formation of photoreceptors in larval and adult goldfish. J. Neurosci. 1982;2(2):178–198. [PubMed]
  • Julian D, Ennis K, Korenbrot JI. Birth and fate of proliferative cells in the inner nuclear layer of the mature fish retina. J. Comp. Neurol. 1998;394(3):271–282. [PubMed]
  • Kay JN, Finger-Baier KC, Roeser T, Staub W, Baier H. Retinal ganglion cell genesis requires lakritz, a Zebrafish atonal Homolog. Neuron. 2001;30(3):725–736. [PubMed]
  • Kennedy BN, Vihtelic TS, Checkley L, Vaughan KT, Hyde DR. Isolation of a zebrafish rod opsin promoter to generate a transgenic zebrafish line expressing enhanced green fluorescent protein in rod photoreceptors. J. Biol. Chem. 2001;276(17):14037–14043. [PubMed]
  • Kennedy BN, Alvarez Y, Brockerhoff SE, Stearns GW, Sapetto-Rebow B, Taylor MR, Hurley JB. Identification of a zebrafish cone photoreceptor-specific promoter and genetic rescue of achromatopsia in the nof mutant. Investig. Ophthalmol. Vis. Sci. 2007;48(2):522–529. [PubMed]
  • Kimura A, Singh D, Wawrousek EF, Kikuchi M, Nakamura M, Shinohara T. Both PCE-1/RX and OTX/CRX interactions are necessary for photoreceptor-specific gene expression. J. Biol. Chem. 2000;275(2):1152–1160. [PubMed]
  • Koulen P, Brandstatter JH, Kroger S, Enz R, Bormann J, Wassle H. Immunocytochemical localization of the GABA(C) receptor rho subunits in the cat, goldfish, and chicken retina. J. Comp. Neurol. 1997;380(4):520–532. [PubMed]
  • Larison KD, Bremiller R. Early onset of phenotype and cell patterning in the embryonic zebrafish retina. Development (Camb., England) 1990;109(3):567–576. [PubMed]
  • Levine EM, Green ES. Cell-intrinsic regulators of proliferation in vertebrate retinal progenitors. Semin. Cell Dev. Biol. 2004;15(1):63–74. [PubMed]
  • Linser PJ, Smith K, Angelides K. A comparative analysis of glial and neuronal markers in the retina of fish: variable character of horizontal cells. J. Comp. Neurol. 1985;237(2):264–272. [PubMed]
  • Loosli F, Staub W, Finger-Baier KC, Ober EA, Verkade H, Wittbrodt J, Baier H. Loss of eyes in zebrafish caused by mutation of chokh/rx3. EMBO Rep. 2003;4(9):894–899. [PMC free article] [PubMed]
  • Luo W, Williams J, Smallwood PM, Touchman JW, Roman LM, Nathans J. Proximal and distal sequences control UV cone pigment gene expression in transgenic zebrafish. J. Biol. Chem. 2004;279(18):19286–19293. [PubMed]
  • Mani SS, Batni S, Whitaker L, Chen S, Engbretson G, Knox BE. Xenopus rhodopsin promoter. Identification of immediate upstream sequences necessary for high level, rod-specific transcription. J. Biol. Chem. 2001;276(39):36557–36565. [PubMed]
  • Masai I, Stemple DL, Okamoto H, Wilson SW. Midline signals regulate retinal neurogenesis in zebrafish. Neuron. 2000;27(2):251–263. [PubMed]
  • Mathers PH, Jamrich M. Regulation of eye formation by the Rx and pax6 homeobox genes. Cell. Mol. Life Sci. 2000;57(2):186–194. [PubMed]
  • Mathers PH, Grinberg A, Mahon KA, Jamrich M. The Rx homeobox gene is essential for vertebrate eye development. Nature. 1997;387(6633):603–607. [PubMed]
  • Morcos PA. Achieving targeted and quantifiable alteration of mRNA splicing with Morpholino oligos. Biochem. Biophys. Res. Comm. 2007;358(2):521–527. [PubMed]
  • Moritz OL, Peck A, Tam BM. Xenopus laevis red cone opsin and Prph2 promoters allow transgene expression in amphibian cones, or both rods and cones. Gene. 2002;298(2):173–182. [PubMed]
  • Mu X, Klein WH. A gene regulatory hierarchy for retinal ganglion cell specification and differentiation. Semin. Cell Dev. Biol. 2004;15(1):115–123. [PubMed]
  • Nasevicius A, Ekker SC. Effective targeted gene ‘knockdown’ in zebrafish. Nat. Genet. 2000;26(2):216–220. [PubMed]
  • Nelson S, Frey R, Wardwell S, Stenkamp D. The developmental sequence of gene expression within the rod photoreceptor lineage in embryonic zebrafish. Dev. Dyn. 2008;237(10):2903–2917. [PMC free article] [PubMed]
  • Nornes S, Clarkson M, Mikkola I, Pedersen M, Bardsley A, Martinez JP, Krauss S, Johansen T. Zebrafish contains two pax6 genes involved in eye development. Mech. Dev. 1998;77(2):185–196. [PubMed]
  • Ochocinska MJ, Hitchcock PF. Dynamic expression of the basic helix–loop– helix transcription factor neuroD in the rod and cone photoreceptor lineages in the retina of the embryonic and larval zebrafish. J. Comp. Neurol. 2007;501(1):1–12. [PubMed]
  • Ohuchi H, Tomonari S, Itoh H, Mikawa T, Noji S. Identification of chick rax/rx genes with overlapping patterns of expression during early eye and brain development. Mech. Dev. 1999;85(1–2):193–195. [PubMed]
  • Otteson DC, Hitchcock PF. Stem cells in the teleost retina: persistent neurogenesis and injury-induced regeneration. Vis. Res. 2003;43(8):927–936. [PubMed]
  • Otteson DC, D’Costa AR, Hitchcock PF. Putative stem cells and the lineage of rod photoreceptors in the mature retina of the goldfish. Dev. Biol. 2001;232(1):62–76. [PubMed]
  • Pan Y, Nekkalapudi S, Kelly LE, El-Hodiri HM. The Rx-like homeobox gene (Rx-L) is necessary for normal photoreceptor development. Investig. Ophthalmol. Vis. Sci. 2006;47(10):4245–4253. [PMC free article] [PubMed]
  • Passini MA, Levine EM, Canger AK, Raymond PA, Schechter N. Vsx-1 and Vsx-2: differential expression of two paired-like homeobox genes during zebrafish and goldfish retinogenesis. J. Comp. Neurol. 1997;388(3):495–505. [PubMed]
  • Postlethwait J, Amores A, Cresko W, Singer A, Yan YL. Subfunction partitioning, the teleost radiation and the annotation of the human genome. Trends Genet. 2004;20(10):481–490. [PubMed]
  • Prabhudesai SN, Cameron DA, Stenkamp DL. Targeted effects of retinoic acid signaling upon photoreceptor development in zebrafish. Dev. Biol. 2005;287(1):157–167. [PMC free article] [PubMed]
  • Raymond PA. Cytodifferentiation of photoreceptors in larval goldfish: delayed maturation of rods. J. Comp. Neurol. 1985;236(1):90–105. [PubMed]
  • Raymond PA, Rivlin PK. Germinal cells in the goldfish retina that produce rod photoreceptors. Dev. Biol. 1987;122(1):120–138. [PubMed]
  • Raymond PA, Barthel LK, Curran GA. Developmental patterning of rod and cone photoreceptors in embryonic zebrafish. J. Comp. Neurol. 1995;359(4):537–550. [PubMed]
  • Rojas-Munoz A, Dahm R, Nusslein-Volhard C. chokh/rx3 specifies the retinal pigment epithelium fate independently of eye morphogenesis. Dev. Biol. 2005;288(2):348–362. [PubMed]
  • Schmitt EA, Dowling JE. Early eye morphogenesis in the zebrafish, Brachydanio rerio. J. Comp. Neurol. 1994;344(4):532–542. [PubMed]
  • Schmitt EA, Dowling JE. Comparison of topographical patterns of ganglion and photoreceptor cell differentiation in the retina of the zebrafish, Danio rerio. J. Comp. Neurol. 1996;371(2):222–234. [PubMed]
  • Schmitt EA, Dowling JE. Early retinal development in the zebrafish, Danio rerio: light and electron microscopic analyses. J. Comp. Neurol. 1999;404(4):515–536. [PubMed]
  • Semon M, Wolfe KH. Reciprocal gene loss between Tetraodon and zebrafish after whole genome duplication in their ancestor. Trends Genet. 2007;23(3):108–112. [PubMed]
  • Shen YC, Raymond PA. Zebrafish cone-rod (crx) homeobox gene promotes retinogenesis. Dev. Biol. 2004;269(1):237–251. [PubMed]
  • Shkumatava A, Fischer S, Muller F, Strahle U, Neumann CJ. Sonic hedgehog, secreted by amacrine cells, acts as a short-range signal to direct differentiation and lamination in the zebrafish retina. Development (Camb., England) 2004;131(16):3849–3858. [PubMed]
  • Stenkamp DL, Barthel LK, Raymond PA. Spatiotemporal coordination of rod and cone photoreceptor differentiation in goldfish retina. J. Comp. Neurol. 1997;382(2):272–284. [PubMed]
  • Stenkamp DL, Frey RA. Extraretinal and retinal hedgehog signaling sequentially regulate retinal differentiation in zebrafish. Dev. Biol. 2003;258(2):349–363. [PubMed]
  • Stenkamp DL, Frey RA, Prabhudesai SN, Raymond PA. Function for Hedgehog genes in zebrafish retinal development. Dev. Biol. 2000;220(2):238–252. [PubMed]
  • Stenkamp DL, Frey RA, Mallory DE, Shupe EE. Embryonic retinal gene expression in sonic-you mutant zebrafish. Dev. Dyn. 2002;225(3):344–350. [PubMed]
  • Summerton J. Morpholino antisense oligomers: the case for an RNase H-independent structural type. Biochim. Biophys. Acta. 1999;1489(1):141–158. [PubMed]
  • Van Raay TJ, Moore KB, Iordanova I, Steele M, Jamrich M, Harris WA, Vetter ML. Frizzled 5 signaling governs the neural potential of progenitors in the developing Xenopus retina. Neuron. 2005;46(1):23–36. [PubMed]
  • Voronina VA, Kozhemyakina EA, O’Kernick CM, Kahn ND, Wenger SL, Linberg JV, Schneider AS, Mathers PH. Mutations in the human RAX homeobox gene in a patient with anophthalmia and sclerocornea. Hum. Mol. Genet. 2004;13(3):315–322. [PubMed]
  • Walther C, Gruss P. Pax-6, a murine paired box gene, is expressed in the developing CNS. Development (Camb., England) 1991;113(4):1435–1449. [PubMed]
  • Wang QL, Chen S, Esumi N, Swain PK, Haines HS, Peng G, Melia BM, McIntosh I, Heckenlively JR, Jacobson SG, Stone EM, Swaroop A, Zack DJ. QRX, a novel homeobox gene, modulates photoreceptor gene expression. Hum. Mol. Genet. 2004;13(10):1025–1040. [PubMed]
  • Wargelius A, Seo HC, Austbo L, Fjose A. Retinal expression of zebrafish six3.1 and its regulation by Pax6. Biochem. Biophys. Res. Comm. 2003;309(2):475–481. [PubMed]
  • Wei Y, Yu L, Bowen J, Gorovsky MA, Allis CD. Phosphorylation of histone H3 is required for proper chromosome condensation and segregation. Cell. 1999;97(1):99–109. [PubMed]
  • Westerfield M. The Zebrafish Book: A Guide for the Laboratory Use of the Zebrafish (Danio rerio) 4th Edition University of Oregon Press; Eugene, Oregon: 2000.
  • Zaghloul NA, Moody SA. Alterations of rx1 and pax6 expression levels at neural plate stages differentially affect the production of retinal cell types and maintenance of retinal stem cell qualities. Dev. Biol. 2007;306(1):222–240. [PubMed]
  • Zhang L, Mathers PH, Jamrich M. Function of Rx, but not Pax6, is essential for the formation of retinal progenitor cells in mice. Genesis. 2000;28(3–4):135–142. [PubMed]
  • Zuber ME, Gestri G, Viczian AS, Barsacchi G, Harris WA. Specification of the vertebrate eye by a network of eye field transcription factors. Development (Camb., England) 2003;130(21):5155–5167. [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • Gene
    Gene
    Gene links
  • Gene (nucleotide)
    Gene (nucleotide)
    Records in Gene identified from shared sequence links
  • GEO Profiles
    GEO Profiles
    Related GEO records
  • HomoloGene
    HomoloGene
    HomoloGene links
  • MedGen
    MedGen
    Related information in MedGen
  • Nucleotide
    Nucleotide
    Published Nucleotide sequences
  • Pathways + GO
    Pathways + GO
    Pathways, annotations and biological systems (BioSystems) that cite the current article.
  • Protein
    Protein
    Published protein sequences
  • PubMed
    PubMed
    PubMed citations for these articles
  • Taxonomy
    Taxonomy
    Related taxonomy entry
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...