Cath6, a bHLH atonal family proneural gene, negatively regulates neuronal differentiation in the retina

Authors

  • Fumi Kubo,

    Corresponding author
    1. RIKEN Advanced Science Institute, Wako, Saitama, Japan
    Current affiliation:
    1. Department of Physiology, University of California, San Francisco, San Francisco, CA 94158-2722
    • Department of Physiology, University of California, San Francisco, 1550 Fourth Street, San Francisco, CA 94158-2722
    Search for more papers by this author
  • Shinichi Nakagawa

    1. RIKEN Advanced Science Institute, Wako, Saitama, Japan
    Search for more papers by this author

Abstract

Basic helix–loop–helix (bHLH) transcription factors play important roles in cell type specification and differentiation during the development of the nervous system. In this study, we identified a chicken homolog of Atonal 8/ath6 (Cath6) and examined its role in the developing retina. Unlike other Atonal-family proneural genes that induce neuronal differentiation, Cath6 was expressed in stem cell-like progenitor cells in the marginal region of the retina, and its overexpression inhibited neuronal differentiation. A Cath6 fused with a VP16 transactivation domain recapitulated the inhibitory effect of Cath6 on neuronal differentiation, indicating that Cath6 functions as a transcription activator. These results demonstrate that Cath6 constitutes a unique member of the Atonal-family of genes in that it acts as a negative regulator of neuronal differentiation. Developmental Dynamics 239:2492–2500, 2010. © 2010 Wiley-Liss, Inc.

INTRODUCTION

During the development of the nervous system, various types of neurons with different functional properties are generated from a pool of progenitor cells. In vertebrate retinas, six types of neurons and one type of glia, i.e., the Müller cells, are generated in a temporally regulated manner (Young,1985; Prada et al.,1991; Stiemke and Hollyfield,1995; Livesey and Cepko,2001). It has been well established that all types of retinal neurons and glia are derived from common multipotent progenitor cells (Turner and Cepko,1987; Holt et al.,1988; Wetts et al.,1989). Multipotent progenitors change their differentiation competence over time and are thereby able to produce a variety of cell types in a well-defined sequence (reviewed in Livesey and Cepko,2001). In contrast to this differentiation phase that takes place in the central region of the retina, the peripheral region of the retina, called the ciliary marginal zone (CMZ), contains stem cell-like progenitor cells and remains undifferentiated for a long period. In a wide range of vertebrate species, cells in the CMZ continue to proliferate until adulthood and generate new neurons, which subsequently incorporate into the pre-existing retina in the central region and undergo differentiation (Hollyfield,1968; Straznicky and Gaze,1971; Johns,1977; Fischer and Reh,2000; Perron and Harris,2000; Moshiri et al.,2004).

The basic helix–loop–helix (bHLH) proteins play important roles in regulating the differentiation competence of multipotent retinal progenitor cells. Several proneural bHLH transcription factors are expressed in the developing retina and control the differentiation of specific cell types in combination with other regulatory factors (Kageyama et al.,1997; Perron and Harris,2000; Vetter and Brown,2001; Hatakeyama and Kageyama,2004). For instance, an Atonal family transcription factor, Ath5, is expressed in retinal progenitors and differentiating retinal ganglion cells (RGCs; Kanekar et al.,1997; Brown et al.,1998; Liu et al.,2001). Forced expression of Ath5 leads to an overproduction of RGCs in various vertebrate species (Kanekar et al.,1997; Fischer and Reh,2000; Matter-Sadzinski et al.,2001), and conversely, this cell type is absent in ath5-mutant animals (Brown et al.,2001; Kay et al.,2001; Wang et al.,2001). Likewise, many other proneural bHLH genes, such as NeuroD, Ath3, Neurogenin2, and an acute-scute family transcription factor, Ash1, are expressed in a highly cell type-specific manner and promote cell type determination and differentiation (Yan and Wang,1998; Morrow et al.,1999; Perron et al.,1999; Tomita et al.,2000; Hutcheson and Vetter,2001; Inoue et al.,2002). These proneural bHLH genes are proposed to be negatively regulated by another class of bHLH genes, Hes genes (Ishibashi et al.,1995; Lee et al.,2005). Hes genes down-regulate proneural gene expression or antagonize its function through protein–protein interactions, resulting in the maintenance of the undifferentiated state of progenitor cells (reviewed in Hatakeyama et al.,2004; Kageyama et al.,2007). In the chicken CMZ, the expression of the Hes family gene, Hairy1, is under the control of Wnt signaling, which is necessary and sufficient for the maintenance of stem cell-like characteristics of the marginal retina, thereby maintaining the undifferentiated state of the CMZ (Kubo and Nakagawa,2009).

We previously isolated a gene fragment that shows a high sequence similarity to Math6 (mouse Atonal homolog 6/Atoh8) during a screening for genes induced upon the activation of Wnt signaling (Kubo and Nakagawa,2009). Here, we cloned a full-length coding sequence of the gene, which we named Cath6 (chicken homologue 6 of atonal). Interestingly, Cath6 was expressed in the stem cell-containing CMZ, unlike the other Atonal-family bHLH transcription factors that are expressed in neurogenic regions. Ectopic expression of Cath6 in the early retinas inhibited the differentiation of RGCs, and this activity was mimicked by its fusion protein with a VP16 transactivation domain, but not by a fusion protein with an Engrailed repressor domain. These results demonstrate that Cath6 is a unique member of the Atonal family of transcription factors that acts as a transcriptional activator to inhibit neuronal differentiation.

RESULTS AND DISCUSSION

Cath6 Is a Novel Atonal Homolog Protein

In our previous study (Kubo and Nakagawa,2009), we obtained a partial cDNA clone that showed a high sequence similarity to the bHLH domain of mouse Atonal homolog 8/Math6 (Inoue et al.,2001) or Hath6 (Wasserman et al.,2002), and this clone was designated Cath6 (chicken homologue 6 of atonal). Sequence comparison between an expressed sequence tag (EST) clone (ChEST 101p14) and a BAC clone containing Cath6 gene locus predicted an open reading frame of 648bp (open boxes, Fig. 1A), which is encoded by three exons. We first examined a total length of Cath6 mRNA by Northern hybridization using Poly A+ RNA isolated from embryonic day (E) 5 chick retinas. Two specific bands were detected corresponding to 2.5 kb and 1.6 kb, respectively (Fig. 1B). To determine if this difference in size of the two transcripts originated from differential polyadenlylation or differential transcriptional start sites, we carried out RNaseH-Northern analysis using the oligonucleotides that hybridize to a region just 3′ to the putative stop codon of Cath6. After the RNaseH digestion, a single 860-bp band was specifically detected (Fig. 1C), suggesting that Cath6 has a single transcriptional start site located 860 bp upstream of the stop codon. We then carried out reverse transcriptase-polymerase chain reaction (RT-PCR) analysis using a set of primers that amplifies the region between the predicted transcription start site and the 5′ end of Cath6 coding sequence (Fig. 1A). As expected, a single band of predicted size was detected (approximately 340 bp, Fig. 1D), whereas the PCR product was not observed using primers designed against genomic sequences further upstream of this region (data not shown). Furthermore, we examined whether this 5′ untranslated region (UTR) is contiguous with Cath6 mRNA by RT-PCR using two sets of primers that amplify the regions between the 5′ UTR and the exon2 or exon3 of Cath6 (Fig. 1A). Specific bands of predicted size were amplified with both combinations of primers (Fig. 1E, approximately 760 bp for 5′ UTR & exon2, and 1,010 bp for 5′ UTR & exon3), indicating that this 5′ UTR is transcribed contiguously with Cath6.

Figure 1.

Molecular characterization of Cath6. A: Schematic representation of the Cath6 gene locus. Two putative transcripts of Cath6, an expressed sequence tag (EST) clone (ChEST 101p14), and a BAC clone (#13-M24) are shown (top). Open boxes and triangles indicate an open reading frame and exon–intron boundaries, respectively. Probe used for Northern analysis is shown as a green line. B: Northern analysis of Cath6 mRNA. C: RNaseH Northern analysis of Cath6 mRNA. Note that a single 0.86-kb band was specifically detected after hybridization with the RNH oligo (shown in Fig. 1A) and RNaseH digestion. An asterisk indicates nonspecific band resulting from RNaseH treatment without oligonucleotides (no oligo). D: Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of 5′ untranslated region (UTR) of Cath6. Note that a specific band was detected when RT+ sample and a plasmid vector containing genomic region of Cath6 (vector) were used as templates, but not in RT− control. E: RT-PCR analysis of Cath6 coding sequences. cDNA fragments of Cath6 were amplified by primers as indicated. F: Comparison of the domain structures of Cath6, Math6, and Net proteins. The number indicates the percentage of amino acid identity. G: Unrooted tree of the chicken Atonal superfamily of proteins and Cash1. The phylogenetic tree is constructed based on the comparison of protein sequences of bHLH domains using the neighbor joining method. Pro, Proline-rich domain; b, basic domain; HLH, helix–loop–helix; Gg, Gallus gallus; Mm, Mus musculus; Hs, Homo sapience; Dm, Drosophila melanogaster.

Sequence analysis of Cath6 genomic locus using NCBI MapViewer (http://www.ncbi.nlm.nih.gov/mapview/) revealed that multiple overlapping EST clones were mapped to the Cath6 locus with two different transcription termination sites (Fig. 1A). These EST clones contained nongenomic poly A sequences and putative poly adenylation signals (AATAAA). Because the distance between the two transcriptional termination sites (840 bp) is consistent with the difference of the two transcripts detected by the Northern blot (2.5 kb and 1.6 kb, Fig. 1B), the two transcripts of Cath6 are presumably derived from differential transcriptional termination.

The full-length cDNA of Cath6 encoded a protein of 218 amino acid residues containing a putative bHLH domain in its carboxyl-terminal region (Fig. 1F; Supp. Fig. S1, which is available online). In the bHLH domain, Cath6 showed 100% identity to both Math6 and Hath6 and 69% identity to Drosophila Net (Brentrup et al.,2000; Fig. 1F). In contrast, no significant homology was found in the amino-terminal region, except for a proline-rich region of low specificity (Fig. 1F). Phylogenetically, Cath6 belonged to a subfamily of atonal-like bHLH factors (Ledent and Vervoort,2001), which formed a group distinct from another proneural gene, Cash1 (Fig. 1G).

Cath6 Is Expressed in the Retinal Stem-like Progenitor Cells at the Ciliary Marginal Zone of the Retina

We next examined the expression pattern of Cath6 in the embryonic retina at different stages by in situ hybridization. At the early optic cup stage (E3.5), strong expression of Cath6 was detected in the peripheral region of the retina (arrows in Fig. 2A), and weak expression was also detected in the central region of the retina and the lens (Fig. 2A), whereas these signals were not detected with sense probes (Fig. 2B). At E5.5, Cath6 continued to be expressed in the less differentiated peripheral retina, but it was not detected in the central retina containing differentiating neural progenitor cells and postmitotic neurons (Fig. 2C). At E7.5, Cath6 was down-regulated in the most peripheral part of the retina and was confined to a region more centrally located (Fig. 2D). The peripheral part of the chicken retina at this stage is subdivided into three regions: the most peripheral part that give rise to the iris/ciliary epithelium, the intermediate region (also referred to as the ciliary marginal zone, CMZ) containing stem cell-like progenitor cells, and the central region that contains undifferentiated neural progenitor cells and differentiating nascent neurons (Moshiri et al.,2004; Kubo and Nakagawa,2009). To further characterize the Cath6-expressing cells, we compared the expression pattern of Cath6 with various molecular markers expressed in the peripheral retina using adjacent sections. The Cath6-expressing region was almost exactly overlapped with the region positive for Rdh10, a marker for the CMZ (Kubo and Nakagawa,2009), but not with the region expressing collagen IX, a marker for presumptive ciliary epithelium/iris (Thut et al.,2001; Kubo et al.,2003; Fig. 2F,G). The Cath6-positive region was also complementary to the regions expressing Notch1, Cash1, and Cath5 (Fig. 2F,G), which are the markers for progenitor cells, preneurogenic progenitor cells, and nascent neurons before terminal differentiation, respectively (Henrique et al.,1997; Matter-Sadzinski et al.,2001; Nelson and Reh,2008). These data suggest that Cath6 is specifically expressed in stem cell-like progenitor cells but not in the iris/ciliary epithelium or in differentiating progenitor cells. Considering that the spatial organization of the peripheral retina represents the temporal order of retinal cell differentiation (Perron et al.,1998), Cath6 presumably starts to be expressed in progenitor cells before the onset of the expression of other proneural bHLH genes that promote neuronal differentiation.

Figure 2.

Expression patterns of Cath6 in the retina. A–D: In situ hybridization of Cath6 in embryonic day (E) 3.5 (antisense probe, A; sense probe, B), E.5.5 (C), and E7.5 (peripheral region, D; central region, E) of chicken retina. Note that Cath6 is specifically expressed in the peripheral retina (arrows in A and C, bracket in D) but not in the central retina. F: The expression of peripheral retinal markers on the serial sections of E7.5 retina. Arrows and arrowheads indicate the putative border between the neural retina/CMZ and the CMZ/ciliary epithelium, respectively. Note that Cath6 expression overlaps with Rdh10 but not with the neurogenic markers, such as Notch1, Cash1, and Cath5 nor with the ciliary epithelium marker collagen IX. G: Schematic representation of the expression patterns of the marker genes in chicken peripheral retinas. Le, lens. Scale bars = 100 μm in A,B,E,F, 200 μm in C,D.

Math6 has previously been shown to be expressed in a wide range of cell types in the central region of mouse retina (Inoue et al.,2001), but no information is available for the expression in the peripheral retina. We thus re-examined Math6 expression in the retina of E18.5 mouse embryos, at a stage comparable to E7.5 chicken retina (Young,1985; Prada et al.,1991), especially focusing on the peripheral region. As previously reported (Inoue et al.,2001), strong Math6 expression was observed in the retinal ganglion cells and a subpopulation of cells in the neuroblastic layer (Supp. Fig. S2B). In addition, we observed strong expression in the most peripheral part of the retina, which topologically correspond to the CMZ (Supp. Fig. S2A). At P8.5, Math6 was weakly expressed in the most peripheral region of the retina differentiating into the ciliary epithelium (Supp. Fig. S2C), as well as in post mitotic neurons, including retinal ganglion cells and amacrine cells (Supp. Fig. S2D; Inoue et al.,2001).

Cath6 Inhibits Differentiation of the Retinal Ganglion Cells

The aforementioned spatial and temporal expression pattern of Cath6 suggests that it may function to maintain the undifferentiated state of stem cell-like progenitor cells in the CMZ, unlike other Atonal superfamily proneural genes that promote neuronal differentiation (reviewed in Kageyama et al.,1997; Perron and Harris,2000; Vetter and Brown,2001; Hatakeyama and Kageyama,2004). To test this idea, we electroporated E1.5 retina with a plasmid vector that expresses Cath6 under the control of the ubiquitous promoter CAG (Niwa et al.,1991), together with a vector expressing green fluorescent protein (GFP) fused to Histone H2B (H2B-GFP) for visualization of the electroporated cells. Under this condition, strong exogenous gene expression continued for only 2 days after the electroporation, namely, up to E3.5, and decayed thereafter. We thus analyzed the differentiation of RGCs, which are the only neuronal cell types found in the retina of such early stages (Prada et al.,1991). In the control retina electroporated with empty pCAG vector and H2B-GFP, a subpopulation of cells in the vitreal side of the central retina expressed Islet1, a marker for RGCs at this stage (Austin et al.,1995) (Fig. 3A,C–C′). In the retina-overexpressing Cath6, the number of Islet1-positive cells was decreased compared with the control retina (Fig. 3B,D–D′). For quantitative analyses, we counted the total number of Islet1-positive cells on all the sections throughout the retinas. At E3.5, the number of Islet1-positive cells was considerably variable between individual embryos depending on the subtle differences in their embryonic stages. We therefore scored the ratio of Islet1-positive cells on the electroporated side of the retinas compared with the nonelectroporated side of the same embryos because the electroporation was unilaterally performed on the right side, leaving the left side of the same embryos a stage-matched control. In Cath6-overexpressing embryos, we found an approximately 60% decrease in the ratio of Islet1-expressing cells compared with the control embryos (P < 0.01, Fig. 3E). On the other hand, mutant Cath6 (Cath6 Δbasic) lacking the basic region of bHLH domain, which is required for its DNA-binding activity (Bertrand et al.,2002), did not decrease the ratio of Islet1-positive cells (Fig. 3E).

Figure 3.

Overexpression of Cath6 inhibits retinal ganglion cell differentiation. A–D′: Effects of Cath6 overexpression on retinal ganglion differentiation. A,B: Embryonic day (E) 1.5 optic vesicles were electroporated with empty pCAG vector (A) or pCAG-Cath6 (B) together with Histone H2B-green fluorescent protein (GFP), and stained with anti-GFP and DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride; A,B) and Islet1 (C,C′,D,D′) at E3.5. C′ and D′: The enlarged views of the central retinas of C and D, respectively. E: The ratio of the number of Islet1-positive cells on the electroporated (EP) side to that of the control side from the same embryos. More than 10 embryos were examined for each construct. Cath6 inhibited the differentiation of Islet1-positive RGCs, whereas Cath6 Δbasic had no effects. Asterisks indicate a statistical difference against the control. **P < 0.01. n.s., not significant (P > 0.05). Scale bars = 100 μm in A–D, 50 mm in C′,D′.

To further characterize the effect of Cath6 overexpression, we examined the expression of other RGC markers, Hu C/D (Fischer and Reh,2000), and neurofilament (NF; Austin et al.,1995; McCabe et al.,1999). As previously reported, Hu C/D was expressed in mature RGCs in the vitreal side of the central retina, and NF was expressed in newly differentiated ganglion cells at ventricular surface and migrating ganglion cells as well as those in the vitreal surface of the retina in the control embryos (Fig. 4A,C,E,G). In Cath6-expressing retina, the number of Hu C/D and NF-positive RGCs were both reduced compared with the control (Fig. 4B,D,F,H). The quantitative analyses showed that the ratio of Hu C/D and NF-positive cells on the electroporated side over control side was significantly smaller in Cath6-expressing retina compared with the control (Fig. 4K; P < 0.01 for Hu C/D, P < 0.03 for NF), suggesting that Cath6 overexpression also inhibits other mature RGC marker expression.

Figure 4.

Effect of Cath6 overexpression on various retinal ganglion cell markers. A–H: Embryonic day (E) 1.5 optic vesicles were electroporated with empty pCAG vector (A) or pCAG-Cath6 (B) together with Histone H2B-green fluorescent protein (GFP), and stained with anti-GFP and DAPI (A–D), Hu C/D (E,F), and neurofilament (NF) (G,H) at E3.5. C and D show the enlarged views of the central retinas of A and B, respectively. I,J: Expression of Cath5 in control (I) or Cath6-overexpressing retina (J). Cath5 expression was not obviously affected by Cath6 overexpression. K: The ratio of the number of Hu C/D or NF-positive cells on the electroporated (EP) side compared with that of the control side. More than 10 embryos were examined for each construct. Asterisks indicate a statistical difference against the control. *P < 0.03, **P < 0.01. Scale bars = 100 μm in A,B; 50 mm in C–J.

These observations suggest that Cath6 inhibits RGC differentiation; however, it is also possible that Cath6 maintains retinal cells in a RGC precursor state during RGC differentiation and thereby prevents their terminal differentiation. We therefore examined the expression of Cath5, which starts to be expressed in RGC precursor cells well before the expression of mature RGC-specific markers (Fischer and Reh,2000). As previously reported (Liu et al.,2001; Matter-Sadzinski et al.,2001), Cath5 was expressed in the migrating RGC progenitor/precursor cells in the central retina (Fig. 4I). Cath6 overexpression did not change the expression pattern of Cath5, indicating that Cath6 does not increase the number of neural precursor cells (Fig. 4I,J).

We also examined the effect of Cath6 overexpression on other retinal neurons (i.e., photoreceptor, bipolar, amacrine, and horizontal cells), which normally differentiate at much later stages in retinal development than RGCs (Prada et al.,1991). At E3.5, the expression of molecular markers specific to photoreceptor, bipolar, amacrine, and horizontal cells were not detected in Cath6-overexpressing retina or in control retina (data not shown), indicating that Cath6 does not induce premature differentiation of these later-born neurons at the expense of RGCs.

Cath6 Functions as a Transcriptional Activator

The activity of Cath6 to inhibit neuronal differentiation was unexpected because all of the Atonal superfamily molecules thus far identified promote neuronal differentiation (Yan and Wang,1998; Morrow et al.,1999; Perron et al.,1999; Tomita et al.,2000; Hutcheson and Vetter,2001; Inoue et al.,2002) and act as transcriptional activators (Cabrera and Alonso,1991; Johnson et al.,1992; Bertrand et al.,2002). We speculated that Cath6 might function as a transcriptional repressor to counteract the function of these Atonal family genes, although Cath6 did not display significant sequence similarities to known transcriptional repressors. To test this possibility, we generated two types of chimeric constructs of Cath6; one was fused with the transcriptional activation domain of VP16 (Sadowski et al.,1988; Shimizu et al.,2002) and the other one fused with the transcriptional repressor domain of Engrailed (EnR; Fan and Sokol,1997; Shimizu et al.,2002) (Fig. 5A). In the VP16-Cath6–expressing retina, the number of Islet1-positive RGCs was dramatically reduced compared with the control retina (Fig. 5B,C,E,E′,F,F′). It should be noted that this effect was much stronger than that of the full-length Cath6 (Fig. 3B,D,D′). In contrast, EnR-Cath6 overexpression caused the overproduction of the RGCs (Fig. 5D,G,G′). We quantified the total number of Islet1-positive cells on the electroporated side and the control side of the embryos under each condition. In VP16-Cath6–overexpressing retina, the ratio of Islet1-positive RGCs was reduced to approximately 27% of the control (P < 0.01, Fig. 5H), whereas EnR-Cath6 overexpression increased the ratio of Islet1-positive RGCs by 64% compared with the control (P < 0.05, Fig. 5H). These data suggest that Cath6 acts as a transcriptional activator to inhibit RGC differentiation.

Figure 5.

Cath6 acts as a transcriptional activator to inhibit retinal ganglion cell differentiation. A: Chimeric constructs of Cath6 fused with the VP16 transactivation domain (VP16-Cath6) or the EnR repressor domain (EnR-Cath6). BG′: Effects of VP16-Cath6 and EnR-Cath6 on retinal ganglion differentiation. Embryonic day (E) 1.5 optic vesicles were electroporated with control pCAG empty vector (B), pCAG-VP16-Cath6 (C), and pCAG-EnR-Cath6 (D) together with Histone H2B-GFP, and stained with anti-GFP and DAPI (B–D) and anti-Islet1 (E–G,E′–G′) at E3.5. E′, F′, and G′ show the enlarged views of the central retina of E, F, and G, respectively. H: The ratio of the number of Islet1-positive cells on the electroporated (EP) side to that of the control side. More than 10 embryos were examined for each construct. Asterisks in H indicate a statistical difference against the control. **P < 0.01, *P < 0.05. Scale bars = 100 μm in B–G, 50 μm in E′–G′.

Molecular Diversity of Proneural Genes and Functional Implications of Cath6 During Retinal Development

We have thus identified a distinguishing member of bHLH transcription factors of the Atonal superfamily, Cath6. In the retina, all of the Atonal superfamily genes analyzed to date promote the differentiation of specific types of neurons (Yan and Wang,1998; Morrow et al.,1999; Perron et al.,1999; Tomita et al.,2000; Hutcheson and Vetter,2001; Inoue et al.,2002). However, Cath6 inhibits neuronal differentiation, even though it functions as a transcriptional activator like other proneural genes. Cath6 belongs to the Ath6/Net family of the Atonal superfamily, and the coding region of this family of molecules is encoded by three exons in chicken, mouse, and humans (Inoue et al.,2001). Considering that the coding regions of other Atonal superfamily genes are usually encoded by a single exon (Sommer et al.,1996; Tsuda et al.,1998), Ath6/Net family genes may have originated from distinct ancestors and constitute a group of molecules with unique functions.

We found that the DNA binding domain of Cath6 is required for its activity, indicating that the intact HLH domain of Cath6 is not sufficient for its function. Some bHLH genes, such as the Hes or Id families, are reported to antagonize proneural bHLH proteins by forming nonfunctional heterodimers with their partner proteins called E proteins through their HLH domains (reviewed in Kageyama et al.,2005,2007). Our finding excludes a possibility that Cath6 acts through this sequestering mechanism and suggests that Cath6 functions by a different molecular mechanism from the Hes or Id families. Because a constitutively active form of Cath6 (VP16-Cath6) mimicked the effect of Cath6, this result strongly suggests that Cath6 acts as an active transcriptional activator. We therefore propose that Cath6 activates the transcription of currently unknown downstream genes that, in turn, inhibit the differentiation process of RGCs. The identification of the target genes of Cath6 will clarify how Cath6 inhibits neuronal differentiation during retinal development.

We have carried out microRNA (miR) -based RNAi experiments against Cath6 to block its endogenous activity. However, we failed to knockdown the endogenous Cath6 expression in the retina with the miR constructs we generated, presumably due to the dilution of miR vectors resulting from the fast proliferation of the retinal progenitor cells. Thus, we cannot rule out a possibility that the phenotypes induced by Cath6, VP16-Cath6, and EnR-Cath6 result from interference with other Ath family members, rather than the endogenous Cath6. More effective knockdown experiments will be required to elucidate the physiological function of Cath6.

In mice, Math6 is implicated in kidney and pancreas development (Ross et al.,2006; Lynn et al.,2008). In the pancreas, Math6 down-regulates the expression of several key transcription factors such as Neurogenin3 by acting as a transcriptional repressor (Lynn et al.,2008). It remains to be tested whether the difference in transcriptional activities of Cath6 and Math6 depends on their molecular properties or the developmental contexts in which they function. Because the targeted deletion of Math6 in mice results in early embryonic lethality around E8.5 (Lynn et al.,2008), Catth6 may also play an important role in early development.

EXPERIMENTAL PROCEDURES

Cloning of Cath6 Full-Length cDNA

To obtain the full-length cDNA of Cath6, an EST clone (ChEST 101p14) that contains partial sequence of Cath6 was purchased from Geneservice. Because this EST clone did not cover the N terminal end of the Cath6 protein, we isolated a genomic fragment that contains the N terminal region of Cath6 by screening the chicken BAC library (Geneservice). Using a probe containing partial sequences of Cath6, a positive BAC clone (#13-M24, Geneservice) was identified. To isolate the N terminal sequence of Cath6, the BAC clone was digested with EcoRI, electrophoresed, and probed with the sequence derived from the most 5′ end of the Cath6 EST clone, which was predicted to be a putative exon1 from the comparison with mouse sequences. A positive DNA fragment (approximately 2Kb) was isolated and cloned into the EcoRI site of the pBSKS vector (pBSKS C1), and the insert sequence was determined by sequencing. The putative methionine start site was found in a frame located at the 5′ upstream region of the putative exon1 and was preceded by an in-frame stop codon. These sequence data have been submitted to the GenBank database under accession number GU590867.

RNaseH Treatment, Northern Hybridization, and RT-PCR

Northern blot analysis was performed according to a standard protocol using a digoxigenin (DIG) -labeled RNA probe designed for the C terminal region of Cath6 (Fig. 1A). For RNaseH-Northern analysis, 3 μg poly(A)+ RNA from E5 chicken retina was mixed with 25 pmol RNaseH (RNH) oligo (5′-ACTGGCTGGAAGCTCTTC TGAATC-3′) and denatured for 5 min at 65°C, followed by annealing for 10 min at 37°C. They were then treated with RNaseH (Toyobo) for 30 min at 37°C and subjected to Northern blot analysis. RT-PCR was performed as previously described (Kubo et al.,2005) using poly(A)+ RNA isolated from E5 chicken retinas. Plasmid vector containing genomic region of Cath6 (pBSKS C1) was used as a positive control. Primers used are the following: 5′ FW: 5′-GGACTGACGGAC GCACGGAC-3′, 5′ RV: 5′-TCAGCACT TTGAGCTCCACG-3′, ex2 RV: 5′-TCA GCGCTGTAATCCAGGTC-3′, ex3 RV: 5′-CCTTTCTGTTCCAGAGACTG-3′. PCR reaction was carried out under cycle conditions of 98°C for 10 sec, 55°C 30 sec, and 72°C for 1 min for 35 cycles. The sequences of the PCR products were checked by direct sequencing using each of the primers.

Phylogenetic Analysis

The phylogenetic relationship of atonal homolog proteins was determined by comparing amino acid sequences of bHLH domains using the neighbor joining method by CulstalW (Thompson et al.,1994). The phylogenetic tree was constructed using Tree View software (Page,1996). GenBank accession numbers are Gg.Cath6 (GU590867), Mm.Math6 (NP722473), Dm.Net (AAF51562), Hs.Hath6 (NP116216), Mm.Neurogenin3 (NP033849), Hs.Neurogenin3 (NP066279), Gg.Neurogenin2 (NP990127), Mm.Neurogenin2 (NP033848), Hs.Neurogenin2 (NP076924), Gg.Neurogenin1 (NP990214), Mm.Neurogenin1 (NP035026), Hs.Neurogenin1 (NP006152), Mm. NeuroD2 (NP035025), Hs.NeuroD2 (NP006151), Gg.NeuroD1 (NP990251), Mm.NeuroD1 (NP035024), Hs.NeuroD1 (NP002491), Gg.NeuroD4 (also known as Ath3/Atoh3, NP990407), Mm. NeuroD4 (NP031527), Hs.NeuroD4 (NP_ 067014), Gg.Ath1/Cath1 (AAO59913), Mm.Ath1/Math1 (NP031526), Hs.Ath1/ Hath1 (NP031526), Gg.Ath5/Cath5 (NP989999), Mm.Ath5/Math5 (NP058560), and Hs.Ath5/Hath5 (NP660161).

Vector Construction

To obtain Full length cDNA of Cath6, the genomic fragment containing the predicted ATG codon (pBSKS C1) was ligated in-frame with N terminal region of Cath6 encoded in ChEST 101p14 using a SacII site within Cath6 cDNA. This plasmid was then used as a template to PCR amplify the entire coding sequences of Cath6 using the following primers: Cath6 FW: 5′-ATGTCGACGCCACCATGCG GAGCACGCCGCTGGCC-3′, Cath6 RV: 5′-ATAGATCTTTACTCCTTCCTTTTTT TAGA-3′, and then subcloned into Sal I/BglII sites of the pCAG vector (Niwa et al.,1991). Cath6 Δbasic construct was generated by removing the basic region of Cath6 by site-directed mutagenesis. pCAG-EnR-Cath6 and pCAG-VP16AD-Cath6 plasmids were constructed by fusing the Cath6 coding sequence to an EnR domain of pCS2+ EnR or a VP16 trans-activation domain of pCS2+ NLS VP16AD (Shimizu et al.,2002), respectively, at its amino-terminus. All of the constructs were verified by sequencing.

In Ovo Electroporation and Phenotypic Analysis

In ovo electroporation of the retina was carried out as previously described (Kubo et al.,2003). In all experiments, the right side of optic vesicles of E1.5 embryos was electroporated with pCAG expression plasmids (5 μg/μl) together with Histone H2B-GFP–expressing plasmid (pCAG H2B-GFP, 0.5 μg/μl), leaving the left side intact. For control experiments, the same amount of empty pCAG vector was electroporated. For the cell counting of RGCs, we prepared serial sections from the entire retinas of electroporated embryos and scored the total number of RGCs that are found on all the sections (typically 20–30 sections). Because the number of RGC marker-positive cells at E3.5 was considerably variable between individual embryos due to subtle difference of the staging, the total number of RGC marker-positive cells of the electroporated side (right side) was normalized to that of the nonelectroporated side (light side).

Immunohistochemistry and In Situ Hybridization

The antibodies used were the following: mouse anti-Islet1 (clone 39.4D5, DSHB, 1:10), rabbit anti-GFP (#598, MBL, 1:500), mouse anti-Hu C/D (Molecular Probes, 1:500), mouse anti-Neurofilament (NF; clone RMO270, Zymed, 1:1,000), Cy3-conjugated anti-mouse IgG (Chemicon, 1:250), and Alexa 488-conjugated anti-rabbit IgG (Molecular Probes, 1:250). The embryos were fixed for 1 hr at room temperature in 4% paraformaldehyde in phosphate buffered saline (PBS), cryoprotected in 30% sucrose for 1 hr, embedded in OCT compounds, and sectioned at a thickness of 10–15 μm. The sections were then permeabilized in 100% methanol for 5 min at −20°C and processed for the standard immunostaining protocol. For in situ hybridization, embryos were fixed in 4% paraformaldehyde overnight at 4°C and processed for the standard in situ hybridization protocol. Sense and antisense in situ probes of Cath6 were transcribed with T7 and T3, respectively, using an EST clone ChEST 101p14 as a template. To make an in situ probe for Cash1, its 3′ UTR sequence was amplified by RT-PCR using chicken retinal cDNA and subcloned into pCRII (Invitrogen). The Math6 in situ probe was provided by Dr. Kageyama (Inoue et al.,2001). Photomicrographs were obtained with Olympus BX51 microscope and Olympus DP70 digital camera using DP control software (Olympus).

Acknowledgements

We thank Drs. Masahiko Hibi for the pCS2 + NLS VP16AD and pCS2 + EnR plasmids and Ryoichiro Kageyama for the Math6 cDNA and in situ probe. We also thank Drs. Masatoshi Takeichi and Ryoichiro Kageyama for discussions and comments and Kentarou Ishida, Tomoko Inagaki, and Chieko Nashiki for technical assistance. This work was supported by a grant-in-aid for scientific research from the Japan Society for the Promotion of Science (to S.N.). F.K. is a recipient of a Fellowship of the Special Postdoctoral Researchers Program, RIKEN.

Ancillary