Rybp, a polycomb complex-associated protein, is required for mouse eye development
© Pirity et al; licensee BioMed Central Ltd. 2007
Received: 26 September 2006
Accepted: 30 April 2007
Published: 30 April 2007
Skip to main content
© Pirity et al; licensee BioMed Central Ltd. 2007
Received: 26 September 2006
Accepted: 30 April 2007
Published: 30 April 2007
Rybp (Ring1 and YY1 binding protein) is a zinc finger protein which interacts with the members of the mammalian polycomb complexes. Previously we have shown that Rybp is critical for early embryogenesis and that haploinsufficiency of Rybp in a subset of embryos causes failure of neural tube closure. Here we investigated the requirement for Rybp in ocular development using four in vivo mouse models which resulted in either the ablation or overexpression of Rybp.
Our results demonstrate that loss of a single Rybp allele in conventional knockout mice often resulted in retinal coloboma, an incomplete closure of the optic fissure, characterized by perturbed localization of Pax6 but not of Pax2. In addition, about one half of Rybp-/- <-> Rybp+/+ chimeric embryos also developed retinal colobomas and malformed lenses. Tissue-specific transgenic overexpression of Rybp in the lens resulted in abnormal fiber cell differentiation and severe lens opacification with increased levels of AP-2α and Sox2, and reduced levels of βA4-crystallin gene expression. Ubiquitous transgenic overexpression of Rybp in the entire eye caused abnormal retinal folds, corneal neovascularization, and lens opacification. Additional changes included defects in anterior eye development.
These studies establish Rybp as a novel gene that has been associated with coloboma. Other genes linked to coloboma encode various classes of transcription factors such as BCOR, CBP, Chx10, Pax2, Pax6, Six3, Ski, Vax1 and Vax2. We propose that the multiple functions for Rybp in regulating mouse retinal and lens development are mediated by genetic, epigenetic and physical interactions between these genes and proteins.
The vertebrate eye is a complex neurosensory organ. Normal function of the eye requires precise spatial organization and interaction between individual tissues to respect the laws of optics. During embryonic development, reciprocal inducing events result in the formation of the lens and the retina that originate from progenitor cells located in the head surface ectoderm and neuroepithelium of the ventral diencephalon, respectively (see reviews:[1, 2]). Abnormal lens and retinal development can cause isolated or widespread ocular abnormalities that can obstruct vision at different levels and lead to blindness (reviewed in ).
Between embryonic day E9.5 and E11.5 of mouse development, the optic vesicle undergoes dorso-ventral patterning of the neuroepithelium followed by its invagination to form the bilayered optic cup . This process gives rise to the optic fissure, allowing blood vessels originating from the vascular mesoderm to enter the developing eye. By E13.5, the nasal and temporal retina on either side of the choroid fissures fuses around the optic nerve axons and blood vessels. Failure of this closure results in a specific developmental abnormality, called ocular coloboma. Ocular coloboma is often seen in association with severe neurological and/or craniofacial abnormalities  or can develop as an isolated condition .
Lens differentiation at E12.5 is marked by cellular elongation of the lens cells forming the posterior part of the lens vesicle . Differentiating lens fiber cells upregulate expression of various classes of structural proteins including crystallins, intermediate filament bead proteins, and cytoskeletal, membrane, and channel proteins. Abnormal lens fiber cell differentiation disrupts lens homeostasis leading to precipitation of lens proteins resulting in lens opacification and perturbed vision (see review ).
At the molecular level, a significant number of genes involved in the control of eye development also regulate brain development . The most notable classes are homeobox genes such as Lhx2, Otx2, Pax6, Rx and Six3, and the basic helix-loop-helix genes Math5, Neurod1 and Neurog2. In contrast, other genes play specialized roles during lens (e.g., Foxe3, Mab21like1, c-Maf, Pitx3, Sox1 and Hsf4) or retinal (e.g., Chx10, Mab21like2, Six6/Optx2, Vax1, Vax2) development [10, 11]. Several global regulatory genes are also integral to normal ocular development. For example, Brg1 , Snf2h  and Sox2  regulate embryonic development prior to the organ formation. Later, Brg1 and Snf2h are thought to regulate retinal  and lens development , and Sox2 is required for lens placode formation  and optic cup formation . AP-2α [19, 20], pRb  and its partners, the E2Fs  play roles in both retinal and lens differentiation.
Rybp (RING1 and YY1 binding protein) is an evolutionarily conserved protein with a zinc-finger motive that was identified first as an interacting partner for the Polycomb group protein Ring1A . Polycomb group (PcG) proteins function as transcriptional repressors acting part through histone modification, and are believed to be important regulators of organogenesis and cell lineage specification [24, 25]. Recent studies in Drosophila have shown that Drosophila RYBP depends upon PcG proteins to repress transcription, suggesting that Rybp could be classified as a PcG protein itself . Other studies have demonstrated Rybp's interaction with DNA binding transcription factors [27–29] as well as with apoptotic  and ubiquitinated proteins . Although the precise molecular function of Rybp is not yet known, these interactions suggest that Rybp may be a multifunctional developmental regulator. Indeed, we have shown recently that the Rybp is required both for early mouse development and for proper brain formation. Specifically a dose-dependent role in the central nervous system (CNS) for Rybp was uncovered: haploinsufficiency in the subset of embryos caused an exencephalic phenotype due to imperfect closure of the neural tube .
Given established parallels between brain and eye development, we hypothesized that Rybp may also play a role in ocular development. Here we determined Rybp protein localization patterns in the murine eye and analyzed the function of Rybp in four mouse models representing reduced or increased Rybp gene dosage. These studies have shown that aberration in the normal protein levels of Rybp can result in retinal coloboma, abnormal lens and anterior eye development, and corneal neovascularization.
Rybp encodes an essential regulatory protein involved in early post-implantation development of the mouse embryo. In addition, Rybp plays important roles in organogenesis as evidenced by disrupted development of the forebrain found both in Rybp heterozygous null and chimeric embryos . Here we characterized the protein localization and in vivo function of Rybp in another organ – the developing mouse eye. We show that Rybp is required for normal retinal and lens development, and may also be involved in the formation of the anterior eye segment.
Our previous work revealed that during CNS development, Rybp is predominantly localized in postmitotic neurons and differentiated cell types of the developing mouse embryo . In the present study, we show that in the developing mouse eye, Rybp also becomes robustly made in the differentiating layers of the neuroretina including the ganglion and inner nuclear layer cells. This suggests once again a possible role for Rybp in cell cycle exit or the commitment to differentiation. Gaining further insight into the precise cellular role of Rybp during mouse ocular development likely requires a conditional mutant version of the gene where Rybp can be eliminated in specific cell lineages or at specific time points of embryonic development.
Summary of ocular phenotypes
No. with ocular phenotype/No. of examined mutants
Coloboma (by E18.5)
Rybp heterozygous (+/-) mice
Coloboma, lens defects (By E14.5)
Chimeric mice Rybp-/- <->
3 non exencephalic/19
Cataract (by 3 month)
ROSA26-Rybp/EGFP; αA-crystallin/Cre Tg/+ mice
Neovascularization (by 3 month)
ROSA26-Rybp/EGFP; β-Actin/Cre Tg/+ mice
Neovascularization (by 3 month)
ROSA26-Rybp/EGFP; β-Actin/Cre Tg/Tg mice
Coloboma, irregular retinal folds, anterior chamber defects (by P7)
ROSA26-Rybp/EGFP; β-Actin/Cre Tg/Tg mice
Aside from the retinal colobomas, a wide range of lens abnormalities were found in the Rybp mouse models. In both Rybp heterozygous and chimeric eyes, ventral rotation of the lens was found in association with coloboma (see Fig. 3 and 5). A similar phenotype has been reported for Otx mutant mice . Lenses of Rybp chimeric embryos showed abnormal separation of the lens vesicle from the surface ectoderm (Fig. 5), a condition found in Pax6 heterozygous  and Foxe3 homozygous  mouse embryos. Transgenic lenses overexpressing Rybp developed abnormally differentiated lenses at birth (if not earlier), and this condition further deteriorated with age resulting in a total collapse of the lens fiber mass (see Fig. 7). This demonstrates that high level of Rybp disrupts normal fiber cell differentiation and maturation. Similar phenotype was observed when AP-2α transcription factor, individual E2Fs or growth factors were overexpressed in lens specific manner ([19, 45], see review ). Whether excess of Rybp interferes with adhesion and migration (like AP-2α), cell cycle (like E2Fs), terminal differentiation (like TGFβ) of secondary fiber cells or acts via completely different mechanisms, requires further studies. Indeed, RYBP was shown to be involved in specification of function within the family of E2F transcription factors in cell culture systems  however this has not been investigated in vivo during lens fiber cell differentiation.
Ubiquitous overexpression of Rybp caused similar lens defects and additionally led to corneal neovascularization (see Fig. 8). Genes such as VEGF, FGF2 and MMP-2 have already been associated with corneal neovascularization [47–50]. A possible connection between Rybp and angiogenesis warrants further investigation, especially in light of the observation that during embryonic development Rybp is expressed in hyaloid vessels and in the endothelium of blood vessels outside of the eye (Fig. 1; Pirity and Schreiber-Agus, unpublished data).
The variable penetrance of the phenotypes may indicate that expression and/or function of the retained wild type Rybp allele is being modified differentially in affected or non-affected animals. This is also supported by our previous observation that the penetrance of the exencephaly, a previously described heterozygous semipenetrant phenotype, was influenced by genetic background . Indeed the phenomenon "semi-penetrance" is a common observation in genetically-engineered mice.
One potential molecular mechanism for abnormal eye development in Rybp mice could relate to Rybp's role in Polycomb group protein transcriptional regulation . Polycomb proteins function in multiprotein complexes to regulate expression of homeotic and other regulatory genes during embryonic development [51–57] and also in eye development [58, 59]. Notably, in our limited gene expression studies using a candidate approach (Figure 9), several key eye regulators were found to be significantly affected in the Rybp transgenic lenses examined, including Sox2 and AP-2α (see also [60, 18, 19, 61]) as well as were some of the crystallins . Moreover, Pax6 localization in retina and lens appears affected in the Rybp heterozygous null mice (Figure 3). Taken together, our findings suggest that Rybp may be regulating the expression of other developmental regulators, whose altered levels in the Rybp mouse models could be causal to the observed phenotypes. Further studies are necessary to determine whether Rybp directly regulates these (and other) genes, and whether this regulation occurs via the interaction of Rybp with Polycomb group transcriptional repressors. Interestingly, the abnormal anterior eye formation of Ring1-deficient mice, which is exacerbated in compound YY1+/-Ring1-/- mice, is similar to what we have observed in the Rybp mutant mice ( and Fig. 7H). Since both YY1 and Ring1A interact with Rybp and with other class II PcG proteins, it is possible that these proteins functionally interact to orchestrate normal vertebral eye development.
Collectively, the present work provides the first in vivo evidence for a role of the Polycomb group binding protein Rybp in mouse eye development and disease. Further studies need to address the molecular basis of this role and to determine how Rybp functionally relates to other known regulators of ocular processes.
ES cell injections, mouse breeding, husbandry, and genotyping of the Rybp knockout colonies was as previously described . Chimeric embryos were generated by microinjecting Rybp-/- R1 ES cells (129/Sv × 129-Cp; ) into blastocysts derived from wild-type (Rybp+/+) mice (C56/BL6), as described previously . The αA-crystallin/Cre transgenic mice were a kind gift of M.L. Robinson (Columbus, Ohio, USA); . Analysis of the Cre-mediated recombination pattern in the αA-Cre line was performed by mating with the ROSA26 reporter line (Gt(ROSA)26Sortm1Sor) as described . The β-Actin/Cre transgenic mice (FVB/N-Tg(ACTB-cre)2Mrt/J) line MRL  were purchased from Jackson laboratories (Bar Harbor, Maine, USA). Cre mice were crossed with the Rybp transgenic mice (see next) to obtain F1 Rybp transgenic mice (FVB).
Mice were housed in a 12-h light, 12-h dark cycle and maintained in the animal facility at AECOM in accordance with institutional guidelines. Noon of the day the vaginal plug was observed was considered E0.5 of embryogenesis. For genotyping targeted ES cell colonies/mice, EGFP Primers: A, (5'-aagttcatctgcaccaccg-3') and B, (5'-tgctcaggtagtggttgtcg-3') were used as described. The genotypes were determined by PCR analysis on DNA extracted from the tail or yolk sac.
The strategy used to generate a Cre recombinase-mediated conditional ectopic Rybp allele by targeting the ubiquitous ROSA26 locus is shown on Figure 6. The ROSA26 locus (referred to as an R26 knock-in) was targeted with a floxed neo cassette followed by an RYBP/EGFP fusion. To generate the RYBP/EGFP fusion, the Rybp open reading frame (ORF) was amplified with primers A (5'-gcacgtcgaccagcccgtccatgaccatgg-3') and B (5'-ctctggatccgaaagattcatcattcactgc 3') and cloned into pEGFP-N3 with NheI/SalI. The Rybp/EGFP fusion was transferred to pBigT  with NheI/Not1, and then the Rybp/EGFP together with the floxed neomycin cassette was cloned into the pROSA26 vector  with AscI/PacI. Rybp transgenic ES cell lines were generated by electroporating linearized targeting vector into R1 ES cells  as described in 96 clones resistant to G418 (Gibco, 300 μg/ml) were selected, and screened by genomic Southern blot hybridization on DNA digested with EcoRV as previously described . Two targeted clones were injected into C57BL/6 blastocysts and produced germ-line chimeras (ROSA26-RYBP/EGFP mice) ; mice carrying the targeted allele were genotyped by the PCR as described . Male chimeras were mated with ICR females, and their agouti offspring were tested for transmission by tail PCR and blotting. Animals heterozygous for the mutation were bred with corresponding Cre transgenic lines and analyzed for the phenotype. Non-transgenic littermates were used as controls for all experiments. Heterozygous transgenic progeny were mated to maintain the allele. All analyses were performed on a mixed (129 × ICR) background and mutant mice were analyzed in comparison to their wild-type littermates.
All embryos were dissected in PBS, fixed overnight in PBS buffered 4% paraformaldehyde, and paraffin-embedded sections (6 μm) were mounted for staining. For immunohistochemistry, deparaffinized and rehydrated tissue slides were first treated for 30 min with 3% H2O2 to inactivate endogenous peroxidases. After rinsing washing in PBS for 5–10 min, slides were blocked with 10% (w/v) BSA in PBS and then incubated overnight at 4°C with antibodies against Rybp (anti-DEDAF; dilution 1:1000; Chemicon, AB3637; rabbit), Pax2 (dilution 1:200, Babco, PRB276P, rabbit), Pax6 (dilution 1:500, mouse IgG1, DSHB), Nestin (dilution: 1:100, mouse IgG1, DSHB Rat-401), NeuN (dilution: 1:1000, Chemicon MAB377), or TUJ1 (dilution: 1:2000, Sigma T8660). After removing excess antibody, samples were incubated with a 1:400 dilution of biotin-conjugated secondary anti-rabbit (Dako, EO466) or anti-mouse antibodies (Dako, EO433) for 45 minutes at room temperature, washed in PBS, and incubated with avidin-biotinylated enzyme complex for 45 minutes. The reaction was developed with the DAB kit (Vector labs). For better visualization, slides were often slightly counterstained with hematoxylin. Samples were viewed and photographed under epifluorescent illumination with Leica MZFLIII microscope.
Dissected P1 and P21 day old eyes were placed in Ito's fixative  for 24 h after the cornea had been pierced with a fine needle. The eyes were washed overnight in cacodylate buffer, postfixed with OsO4, dehydrated, and embedded in Epon (Roth, Karlsruhe, Germany). Semithin sections were stained with toluidine blue. Ultrathin sections were stained with uranyl acetate and lead citrate and viewed with a Zeiss (Oberkochen, Germany) EM 902 electron microscope.
Western blotting analysis was conducted upon protein extracts from embryonic stem cells as described earlier . Primary antibodies were used against RYBP (anti-DEDAF; 1:1,000, Chemicon AB3637), GFP (1:1,000, Molecular Probes A11122) and Flag HRP (1:5000, M2, Sigma).
Total RNA isolated from the lens and retina of the ROSA26-Rybp/EGFP; αA-crystallin/Cre P1 and P21 day old mice was extracted and reverse transcript was obtained as described elsewhere . Real time PCR was performed as previously described using primers to amplify Pax6, Prox1, MafA, MafB and c-Maf transcripts [69, 70]). The additional primers are: Rybp primers A, (5'-agaccagcgaaacaaaccac-3') and B, (5'-aggaggagcgagtcttttcc-3'); Crystallin βA4 (Cryba4) primers A, (5'-gggtttgttcccagttcct-3') and B, (5'-acctgagtggtgatcgctct-3'); Filensin (Bfsp1) Primers: A, (5'-cattgagattgaaggcagca-3') B, (5'-acactggatccaaggctgag-3'); AP2α primer A, (5'-gtgtcagagatgctgcggta-3') and B (5'-tgaggatggtgtccacgta-3'); Integrin α-6 primers A, (5'-attctcctgagggcttccat-3') and B, (5'-ttgagggaaacaccgtcact-3'); Sox2 primer A, (5'-acttttgtccgagaccgaga-3') and B, (5'-ctccggcaagcgtgtactta-3'); and B2M primers A, (5'-catacgcctgcagagttaagc-3') and B, (5'-gatgcttgatcacatgtctcg-3'). Amplification of the cDNA was performed using 7900 HP Applied Biosystems Real Time PCR machine. Relative fold changes were calculated using CCNI as an internal control as described .
ganglion cell layer
enhanced green fluorescent protein
hematoxylin and eosin
inner nuclear layer
primary lens fiber cells
polycomb group protein
outer nuclear layer
open reading frame
polymerase chain reaction
quantitative reverse transcriptase – polymerase chain reaction
We are grateful to Dr. M.L. Robinson (Miami University, Oxford, OH) for the αA-crystallin/cre transgenic mice. We thank the Histopathology Shared Resource of the Albert Einstein Cancer Center for histology, Radma Mahmood and Kveta Cveklova for superb technical assistance. We thank Dr. J.D. Locker for his kind help during revisions of this manuscript. This work was supported by NIH grants EY12200 and 14237 (AC) and CA92558 (NSA). A.C. is a recipient of the Irma T. Hirschl Career Scientist Award.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.