*Author to whom all correspondence should be addressed.
Regulatory interrelations among topographic molecules CBF1, CBF2 and EphA3 in the developing chick retina
Version of Record online: 25 DEC 2001
Development, Growth & Differentiation
Volume 41, Issue 5, pages 575–587, October 1999
How to Cite
Yamagata, M., Mai, A., Pollerberg, G. E. and Noda, M. (1999), Regulatory interrelations among topographic molecules CBF1, CBF2 and EphA3 in the developing chick retina. Development, Growth & Differentiation, 41: 575–587. doi: 10.1046/j.1440-169x.1999.00462.x
- Issue online: 25 DEC 2001
- Version of Record online: 25 DEC 2001
- Eph kinase;
- retinotectal system;
- topographic molecules;
- winged-helix transcription factors.
It has been shown that topographic expression of two winged-helix transcription factors, CBF1/c-qin and CBF2, and a receptor tyrosine kinase EphA3 (Mek4/Cek4) play important roles in establishing the topographic retinotectal projection map along the rostrocaudal axis. The interrelationship among these topographic molecules in the chick retina was studied during development. The topographic expression of CBF1 and CBF2 preluded the graded expression of EphA3, but their precise expression profiles did not exactly fit together. However, interestingly, CBF1 and CBF2 were properly expressed, together with EphA3, in immortalized cell lines derived from the quail retina, which maintained position-specific characteristics. The expression of another topographic molecule SOHo-1, the sensory organ homeobox-1 transcription factor, was separate from EphA3 expression. Ectopic expression of CBF1 using in ovo electroporation repressed the expression of CBF2, and misexpression of CBF2 influenced the graded localization of EphA3 in the retina, albeit imperfectly. Taken together, it is suggested that retinal cells first begin to express CBF1 or CBF2 according to their topographic positions, generate cellular descendants in which the expression of CBF1 and CBF2 is maintained cell-autonomously, and then establish the nasotemporal gradient of EphA3 under the control of CBF2, although indirect.
Topographic maps with a defined spatial ordering of neuronal connections are a key feature of brain organization. The most widely used model for studies of the formation of topographic maps is the retinotectal projections, where retinal ganglion cell axons innervate the tectum according to the spatial order determined by their respective positions in the retina ( Cowan & Hunt 1985; Tessier-Lavigne & Goodman 1996). Since Sperry (1963) proposed the chemoaffinity theory in which graded arrangements of molecules in the retina and in the tectum were suggested to confer positional addresses and control the topographic targeting of retinal axons, numerous topographic molecules asymmetrically distributed in the retina and in the tectum have been reported in various animals ( Sanes 1993; Kaprielian & Patterson 1994; Harris & Holt 1995; Roskies et al. 1995 ).
We showed that in the developing chick retina, two winged-helix molecules, CBF1 (also known as c-qin) and CBF2, are expressed in the nasal and temporal parts, respectively, and their ectopic expression in the retina perturbed the topographic order of the retinotectal projection along the rostrocaudal axis ( Yuasa et al. 1996 ). These winged-helix transcription factors are known to be responsible for gene regulatory mechanisms in embryonic development and cellular differentiation (for review, Kaufmann & Knochel 1996). EphA3 (previously called Mek4, Cek4, Hek4, Hek, or Tyro4) is a member of the cell-surface Eph tyrosine kinase family that is involved in various aspects of neural development ( Drescher 1997; Flanagan & Vanderhaeghen 1998), and is topographically expressed in a high temporal-low nasal gradient in the retina ( Cheng et al. 1995 ). This Eph kinase has been hypothesized to be instrumental in the repulsive guidance of the temporal retinal axons in the tectum by detecting reciprocal gradients of their ligands ephrin-A2 (ELF1) and ephrin-A5 (AL-1/RAGS) along the rostrocaudal axis of the tectum ( Drescher et al. 1995 ; Nakamoto et al. 1996 ; Monschau et al. 1997 ; Frisen et al. 1998 ). In the tectum, the graded distributions of these ephrins appear to be arranged by early graded expression of the homeobox transcription factors en-1 and en-2 (Drosophila engrailed homologs; Logan et al. 1996 ; Shigetani et al. 1997 ), suggesting a regulation cascade of topographic molecules in the tectum (see also Rowitch et al. 1997 ; Funahashi et al. 1999 ). However, the relationships among expression of the topographic molecules in the retina remain unclear.
To address this question, we studied the links between CBF1, CBF2 and EphA3 by examining their minute expression patterns during eye development and in immortalized retinal cell lines, and assessed the effects of experimentally misexpressed CBF1 and CBF2 using an in vivo electroporation method on their normal expression patterns. The results obtained suggested a complex regulatory relationship among these molecules during development.
Materials and Methods
Probe preparation and in situ hybridization
Digoxigenin-labeled riboprobes for CBF1, CBF2 and EphA3 were synthesized from NotI- or XhoI-digested cDNA templates in pBluescript KS (–) with T7 or T3 RNA polymerase. The templates used were the 1883 bp fragment (nucleotide residues; – 411–1472) of chick CBF1, the 716 bp fragment (856–1571) of CBF2 ( Yuasa et al. 1996 ) and the 787 bp fragment (865–1651) of chick EphA3 ( Sajjadi et al. 1991 ). Nucleotide residues are numbered in the 5′ to 3′ direction, beginning with the first residue of the initiative methionine. When we previously used the whole sequence of CBF2 including the winged-helix region as a probe for in situ hybridization ( Yuasa et al. 1996 ), not only the temporal retina and diencephalon but also the mesenchyme surrounding the optic tectum showed positive signals. In the present study, we used a probe covering the 3′-half to exclude the mesenchymal staining, which was probably caused by cross-hybridization with other winged-helix molecules (compare Fig. 2 of Yuasa et al. 1996 with Fig. 1H in the present report). Sensory organ homeobox-1 (SOHo-1) cDNA covering the whole coding region (793 bp, – 1–792; Deitcher et al. 1994 ) was prepared by reverse transcription–polymerase chain reaction (RT-PCR) amplification using specific primers and cloned into a plasmid SLAX 12 Nco, a derivative of pBluescript KS(–), ( Morgan & Fekete 1996). The digoxigenin-labeled riboprobe was synthesized from the NcoI- or EcoRI-digested cDNA template with T3 or T7 RNA polymerase.
White leghorn chicken eggs were obtained from a local supplier, and incubated at 37.5°C. Embryos were staged according to Hamburger and Hamilton (1951). Whole-mount in situ hybridization was carried out as described by Wilkinson (1992), and 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (BCIP/NBT) and Fast Red TR/naphthol AS-MX phosphate (4-chloro-2-methylbenzenediazonium/3-hydroxy-2-naphthoic acid 2,4-dimethylanilide phosphate; Sigma Chemical Co., St Louis, MO, USA) were used as alkaline phosphatase substrates for dark-blue and magenta signals, respectively.
Expression plasmids and antibodies
The entire sequences of CBF1 and CBF2 that had been cloned in the Clal site of RCAS-BP(B), ( Yuasa et al. 1996 ), were cloned into the unique Clal site of pMiwClaI, which was prepared from pMiwCAT ( Kato et al. 1990 ) by modifying its HindIII and HpaI sites to a ClaI site using an adaptor (CATCGATG). To construct pMiwSOHo-1, the ClaI fragment in SOHo-1/pSLAX (see above) was cloned into the ClaI site of pMiwClaI. The pMiwCAT and pMiwZ were obtained from Dr H. Kondoh, Osaka University, Japan. Production of CBF1 and CBF2 proteins following transfection of fibroblasts with pMiwCBF1 and pMiwCBF2 was confirmed with monoclonal antibodies to these proteins. Production of SOHo-1 mRNA after in ovo electroporation with pMiwSOHo-1 was observed by in situ hybridization.
To generate hybridomas producing monoclonal antibodies to CBF1 and CBF2, 50 μg each of multivalent antigenic peptide (MAP)-derivatives (Sawady Co., Tokyo, Japan) of a CBF1 sequence (RSTTSRAKLAFKRGARLTSTGLT, amino acid residues 238–260) and a CBF2 sequence (RRKRFKRQQLPAPELLLRAVDPA, 235–257) was mixed with 1 μg of Gerbu adjuvant (Gerbu Biotechnik, Gaiberg, Germany), injected four times at 3 day intervals into footpads of 4-week-old female BALB/c mice. On the next day of the final immunization, the popliteal and inguinal lymph nodes were dissected, and dissociated using a pair of frosted-end microscope slides, and then fused with PAI myeloma cell line (developed by Dr T. Stahelin) with polyethylene glycol by a standard method ( Hockfield et al. 1993 ). Hybridomas were then selected in a culture medium containing hypoxythantine/aminopterine/thymidine and BriClone supplement (BioResearch Ireland, Dublin, Ireland), and screened for each MAP compound by enzyme-linked immunosorbent assay. Positive clones were further screened for immunostaining pMiwCBF1- or pMiwCBF2-transfected quail QT6 fibroblasts, and two resulting clones, SF-CBF1 and SF-CBF2, were used as hybridomas producing monoclonal antibodies to CBF1 and CBF2 proteins, respectively. Rabbit polyclonal antibody to EphA3 (L18, catalog #SC-920) was purchased from Santa Cruz Biotechnology, Santa Cruz, CA, USA). Texas red- conjugated secondary antibodies were from Amersham-Pharmacia, Tokyo, Japan.
In ovo electroporation
Electroporation of chick embryos in the egg shell originally developed by Muramatsu et al. (1997) was modified using an electroporation apparatus consisting of an electro cell manipulator ECM2001 and a graphic pulse opitimizer 500 (BTX, San Diego, CA, USA). Chicken eggs were incubated for 30–40 h to Hamburger-Hamilton (H-H) stage 8 (3–5 somites), rinsed with 70% ethanol, 5 mL of albumen was removed using an 18-gauge needle fitted to a 10 mL syringe, and a window 15 mm in diameter was opened in the shell. In some cases, 10% India ink in Hanks’ balanced salt solution (HBSS) was injected beneath the embryos. Platinum electrodes (5 mm length, 1 mm diameter), specially manufactured by Unique Medical Imada (Sendai, Japan), were held with a micromanipulator on the stage under a binocular microscope. The anode and cathode were placed on the vitelline membrane parallel to the head–trunk axis of the embryo near the anterior neural fold (cf. Fig. 3A), and the area covering the electrodes and embryo was immersed in 10 μL of HBSS.
Plasmid DNA prepared using a plasmid preparation kit (Qiagen, Chatsworth, CA, USA) were suspended at a concentration of 2 μg/μL in 0.05%(w/v) Fast Green (Sigma Chemical Co.), 10 m M Tris-HCl, 1 m M ethylenediamine tetraacetic acid (EDTA), pH 8.0, and injected into the anterior neuropore of the embryo with a pulled glass pipette until the space was filled with the colored solution. To determine the sites of misexpression, pMiwCBF1 or pMiwCBF2 was mixed and cotransfected with pMiwZ as a reporter at a ratio of 10:1. The injected embryos were immediately electroporated four times with a 30 V rectangular pulse for a duration of 60 ms. The electrodes were then removed from the eggs, which were sealed with tape and returned to the humidified incubator until the time of fixation.
Transfection of pMiwCBF1 or pMiwCBF2 per se did not appear to influence the morphology of the embryos, including the eye and brain. In about one-tenth of the embryos, however, small eyes developed, often associated with malformation of the forebrain, probably because of damage caused by electroporation itself as these abnormalities occurred independently of the DNA constructs used. To avoid such artefacts, we analyzed only embryos at E4 with a properly transfected retina and with normal size and morphology. As only the anode-side retina was transfected with plasmids, the other side of the retina in the same embryo was examined as a control for the normal level of endogenous gene expression.
Histochemistry and image analysis
Transfected embryos were dissected at appropriate stages, immersion-fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) at 4°C for 3 h, stained with X-gal (5-bromo-4-chloro-3-indolyl-β- D-galactopyranoside) at room temperature for 2 h as described previously ( Gray & Sanes 1991), and re-fixed with 4% paraformaldehyde/PBS. Whole-mount in situ hybridization using digoxigenin-labeled probes was carried out as described above, but Fast Red TR/naphthol AS-MX phosphate was used for color development in combination with X-gal histochemistry. Hybridization and color reactions in the same series were stopped at the same time. Some embryos were dipped in a series of 12, 15 and 18% sucrose-containing HBSS, mounted in Tissue-Tek (Miles, Elkhart, IN, USA), and cryosectioned. The embryos were photographed under an Olympus microscope equipped with a high-resolution digital camera (RD175; Minolta, Tokyo, Japan), and images were processed using Adobe Photoshop (version 3.0 J) on Power Macintosh computers.
Ribonuclease protection assay for cell lines
Immortalized quail cell lines, RNF5 and RTC5, were maintained in 1:1 Dulbecco modified Eagle medium/ Ham F12 supplemented with 10% fetal calf serum, 1% chicken serum, 0.25% Matrigel (Collaborative Biomedical, Bedford, MA, USA), and 10–5Mβ-estradiol (DH10CMB) at 41.5°C in 7% CO2 incubators ( Pollerberg et al. 1995 ). For differentiation to retinal ganglion cells, 10 ng/mL human recombinant basic fibroblast growth factor (Promega, Madison, WI, USA) was added to the culture medium.
Ribonuclease protection assays (RPA) were carried out using an RPA II ribonuclease protection assay kit (Ambion, Austin, TX, USA). Quail counterparts of CBF1 (corresponding to chick homolog nucleotide residues 883–1120: GenBank accession number U47275, the nucleotide residue at the initiative methionine is + 1), CBF2 (corresponding to chick 1316–1520: U47276), EphA3 (corresponding to chick 104–674: M68514), and glyceraldehyde-3-phosphate dehydrogenase (quail 226–738: Z19086) were amplified from total RNA with E6 quail retina as a template (Titan One tube RT-PCR system; Boehringer Mannheim, Tokyo, Japan). Fragments obtained were cloned into a pGEM-T vector (Promega, Madison, WI USA). The plasmids were digested with either NotI or NcoI in the polylinker sites, except for the plasmid for EphA3, which was digested with AvaI Y, and used for preparation of labeled riboprobes with [α-32P]-UTP using T7 or SP6 RNA polymerase. The probes were hybridized to 5 μg each of total RNA, which was prepared from tissues or cell lines with an Ultraspec RNA reagent (Biotecx, Houston, TX, USA). After ribonuclease treatment, protected RNA was analyzed by 6% polyacrylamide denaturation gel electrophoresis according to the manufacturer’s protocol (Ambion).
Spatiotemporal expression of CBF1, CBF2, SOHo-1, and EphA3 in chick eyes during early developmental stages
Although previous studies revealed the expression patterns of CBF1, CBF2 ( Hatini et al. 1994 ; Shimamura et al. 1995 ; Yuasa et al. 1996 ), and EphA3 ( Cheng et al. 1995 ; Kilpatrick et al. 1996 ; Connor et al. 1998 ) using rodents or chicks, their spatial and temporal expression patterns in the optic vesicle or retina were not fully investigated coherently in a single animal species. Here, we examined their expression during early development in the chick retina by whole-mount in situ hybridizations.
The CBF1 first showed weak expression around the anterior-most neural fold in H-H stage 8–9 (about embryonic days [E]1.5) embryos ( Fig. 1A), which subsequently became strong expression at the anterior prosencephalon in stage 10 embryos ( Fig. 1B). At stage 8–9, no CBF2 or EphA3 expression was detected at the prosencephalic region of the embryos although both were already expressed in the somites (data not shown). At stage 10, weak expression of CBF2 became visible at the ventral-posterior part of the lateral prosencephalon, corresponding to the primary optic vesicle ( Fig. 1E), however, still no EphA3 expression was seen in the prosencephalon ( Fig. 1I). A weak expression of EphA3 in the optic vesicle began at stage 11, where it was homogeneous throughout the vesicle ( Fig. 1J,K). At this stage, CBF1 and CBF2 were topographically expressed in the optic vesicle, each concentrated in about one-third of the entire eye; in the anterior or posterior part, respectively (from stage 11 onward; Fig. 1C,D,F–H). Expression of these winged-helix transcription factors declined as development proceeded, and had almost completely disappeared at stage 36 (E10; Yuasa et al. 1996 ). From stage 15 onward, the earlier non-topographic expression of EphA3 at stages 11–12 changed to a clear temporal-high and nasal-low gradient ( Fig. 1L,M), and the topographic expression continued up to stage 34 (E8; Cheng et al. 1995 ).
In summary, there seemed to be three developmentally distinctive stages with respect to topographic distribution of these three transcripts. At the initial stage (≤ stage 10), the topographic expression of CBF1 and CBF2 in the primary optic vesicle was developmentally linked to their expression in the prospective telencephalon and diencephalon, respectively ( Fig. 1N). At the second stage (stages 12–13), EphA3 was expressed weakly throughout the whole of the optic vesicle ( Fig. 1O). Finally (from stage 15 onward), EphA3 established a nasotemporally graded expression pattern ( Fig. 1P). The expression domains were distinct when strictly observed: CBF1 and CBF2 expression domains did not overlap with each other and were restricted to the anterior and posterior thirds of the optic vesicle, respectively. EphA3 was expressed in a gradient that overlapped with, but was different from, the polarized, non-graded expression of CBF2 ( Fig. 1P).
In addition to these three transcripts, we examined the expression pattern of SOHo-1, as this transcription factor was also reported to be asymmetrically expressed along the nasotemporal axis of the retina ( Deitcher et al. 1994 ). The initial expression of SOHo-1 was observed at stage 11 ( Fig. 2A), and the nasal-high and temporal-low expression was established at stage 12 ( Fig. 2B), when the graded expression of EphA3 was still not clear (see Fig. 1). The expression pattern was still roughly maintained at E4, the latest stage that we examined. However, when carefully analyzed, the expression of SOHo-1 began to diminish at stage 15 at the central nasal portion of the retina ( Fig. 2C). At later stages, we always observed that the central nasal part of the retina had much less expression of SOHo-1 compared with the ventral and dorsal parts ( Fig. 2D–F). At the ventral part, SOHo-1 is expressed also in the temporal part close to the optic fissure ( Fig. 2E,F), suggesting that expression also arranged along the ventral and dorsal axis of the retina rather than a simple gradient along the nasotemporal axis. Thus, the topographic expression of SOHo-1 was clearly distinct from that of CBF1, and not complementary to that of either CBF2 or EphA3, when all of them established unique expression patterns along the nasotemporal axis.
Gene transfer to the developing retina by in ovo electroporation
We previously employed RCAS-BP, a replication- competent retroviral vector, to infect embryos at stages 9–11 and misexpress exogenous genes in the developing chick retinae ( Yuasa et al. 1996 ). This misexpression system using replication-competent RCAS-BP retroviral vector, however, has some restrictions (see also Morgan & Fekete 1996). First, exogenous genes begin to be expressed after one cell cycle postinfection, usually after more than 12 h, as integration of the retroviral vector into the host genome requires mitosis. The early expression of exogenous genes at 12 h postinfection is expected to be limited to a small number of cells, although the infection finally spread widely in a few days postinfection. Second, it is difficult to insert a reporter gene together, as the packagable size in replication-competent retroviral vectors is quite limited (< 2 kb). Third, the promoter for exogenous gene expression in the retroviral vectors is likely not to be strong enough to change cell fate, at least in some tissues.
In the present study, we took advantage of the in ovo electroporation method to overcome these points ( Muramatsu et al. 1997 ; Funahashi et al. 1999 ) to obtain strong expression of exogenous genes selectively at H-H stage 11 in the optic vesicle, as this is the stage when the positional phenotype of retinal ganglion cells is thought to be determined ( Dütting & Thanos 1995; Thanos et al. 1996 ). The exogenous genes inserted in the vector pMiw were expressed under the control of Rous sarcoma virus enhancer and chicken β-actin promoter suitable for strong expression in a wide range of cell types ( Kato et al. 1990 ). Following microinjection of plasmids into stage 8 embryos, about 12–15 h before stage 11, so as to attain a substantial gene expression at stage 11, electroporation was carried out in one direction to transfect only the anode side of the embryos ( Fig. 3A). In typical experiments using pMiwZ (pMiw carrying lacZ), about 60% of the embryos survived until E4 and about 20% survived until E6, but it was difficult to obtain E8 embryos. In many cases, expression was preferentially restricted to the retina at E4, but other regions including retinal pigment epithelium, diencephalon, and telencephalon were also occasionally positive for expression ( Fig. 3B) because of the difficulty in controlling the transfection area. In the present study, we could not investigate the retinotectal projection in electroporated embryos because very few of the embryos survived beyond E8.
Cotransfection of pMiwZ and pMiwCBF1 (or pMiwCBF2) was assessed by double color detection using X-gal histochemistry for lacZ (blue staining) and in situ hybridization with Fast Red/naphthol (magenta staining) as a substrate for alkaline phosphatase. The misexpressed transcripts of CBF1, CBF2 and β-galactosidase were detected at 6 h after electroporation (data not shown) and were strongly present at 12 h when the embryos developed to stage 11 ( Fig. 3C). The transfected cells, however, were sometimes distributed in a punctate pattern in the tissue (see Fig. 3E) as observed also in the retroviral vector-mediated transfection ( Yuasa et al. 1996 ). The plasmid-derived transcripts were detected clearly at E4 (e.g. Fig. 3D,E) and E6 (data not shown). The red ectopic signals derived from an exogenous gene were closely associated with the blue signals ( Fig. 3E–G), showing that co- introduced β-galactosidase works as an indicator for the transfected areas. However, importantly, the lacZ staining was not always associated with the signal for the second exogenous gene at the cellular level ( Fig. 3E–G), indicating that lacZ expression does not precisely reflect the expression of the second exogenous gene.
Effects of misexpressed CBF1 and CBF2 on endogenous CBF2 and CBF1 expression
We examined the effects of misexpressed CBF1 and CBF2 on the endogenous expression of CBF2 and CBF1, respectively. When CBF1 was ectopically expressed either in the temporal region or in the whole of the eye, the level of endogenous CBF2 expression in the temporal area was decreased at E4 (four embryos/10 embryos; Fig. 4A–C,D–F), with different levels of lacZ expression. In these affected embryos, the untransfected contralateral retina maintained the normal expression pattern of CBF2 ( Fig. 4G–I). As a control, lacZ misexpression did not affect the expression of CBF2 (10 embryos; Fig. 4 J–L). When CBF1 was overexpressed only in the nasal retina where endogenous CBF1 was localized, the temporal expression of CBF2 was not affected (two embryos; not shown).
We also introduced pMiwCBF2 into the retina using the same technique to determine whether ectopic CBF2 influences the nasal expression of CBF1. Misexpressed CBF2 did not affect the nasal expression of CBF1 (20 embryos; Fig. 4M–O).
Effects of misexpressed CBF1 and CBF2 on the expression pattern of EphA3
We then examined whether the ectopic expression of CBF1 or CBF2 affected the graded expression of EphA3. When CBF1 was misexpressed in the temporal retina ( Fig. 5A–C), whole retina ( Fig. 5D–F) or nasal retina (data not shown), the overall gradient of EphA3 was not changed in any of the 31 embryos inspected at E4.
On the contrary, when CBF2 was ectopically misexpressed ( Fig. 5G–Q), the expression pattern of EphA3 was markedly changed in some embryos (four embryos/46 embryos). Two embryos in which CBF2 was misexpressed at the ventro-nasal area showed a shift in the domain of the highest level of expression of EphA3 from the temporal to the dorsal region of the retina ( Fig. 5G–I). In another case where misexpression was diffusely distributed throughout the retina, ectopic expression of EphA3 was observed in the nasal part of the retina together with normal temporal expression ( Fig. 5J–L). Misexpression of CBF2 in the telencephalon enhanced expression of EphA3 in this region ( Fig. 5M–O). In the cases shown in Fig. 5(P,Q), ectopic or overexpression of CBF2 was found in the nasal or temporal parts of the eyes, respectively, but any alteration in the expression of EphA3 was not observed. As a control, transfection with pMiwZ did not influence the expression pattern of EphA3 (20 embryos; not shown).
Heritable expression of CBF1 and CBF2 in immortalized quail cell lines
To further analyze the interrelationship between the expression of retinal topographic molecules and the axonal phenotype at the cellular level, we next used v-myc-immortalized cell lines generated from E3.5 embryonic quail retina ( Pollerberg & Eickholt 1995). Cells are homogeneous with respect to positional identity as they are each derived from a single founder cell, and maintain their position-specific phenotype for target specificity ( Pollerberg et al. 1995 ). These cell lines showed rapid and substantial differentiation (≈ 30% of the total cells at most) to the retinal ganglion cell-like cells by addition of basic fibroblast growth factor (bFGF). The axonal processes of the temporal retina-derived cell lines avoided stripes of membrane vesicles from the posterior tectum whereas those of the nasal retina-derived cell lines did not, that is, they displayed a position-dependent target preference in vitro as primary retinal ganglion cells ( Pollerberg & Eickholt 1995).
To examine the expression of CBF1, CBF2, and EphA3 in these cell lines with and without bFGF, we carried out ribonuclease protection assays using specific riboprobes ( Fig. 6A). A nasal retina-derived cell line RNF5 and temporal retina-derived cell line RTC5 specifically expressed the quail counterparts of chick CBF1 and CBF2, respectively. In addition, expression of EphA3 in RTC5 was significantly higher than that in RNF5 expectedly, also at the protein level ( Fig. 6B). These results showed that these quail cell lines maintain not only their characters in the target preference but also the expression profile of the topographic molecules of their originated positions in a heritable manner and autonomously in the in vitro environment.
Early expression of CBF1 and CBF2, and regulatory link between CBF1 and CBF2
We previously showed that winged-helix transcription factors, CBF1 and CBF2, expressed in the nasal and temporal retina, respectively, control the area-specific axon guidance and/or target recognition of the retinal axons ( Yuasa et al. 1996 ). Both of the transcription factors begin to be expressed topographically prior to stage 11 in vivo ( Fig. 1; Yuasa et al. 1996 ) at which stage the nasotemporal fate of retinal ganglion cells was shown to be determined ( Dütting & Thanos 1995; Thanos et al. 1996 ; also see Müller et al. 1998 ). During early development of the nervous system, CBF1 functions as a regulator of cell proliferation and differentiation in the anterior neural tube that gives rise to the telencephalon and anterior optic vesicles ( Hatini et al. 1994 ; Xuan et al. 1995 ). The findings that diffusible factors including FGF8 and BMP4 control telencephalic expression of BF1, a murine ortholog of CBF1 ( Furuta et al. 1997 ; Shimamura & Rubenstein 1997), suggest that the initial expression of CBF1 is triggered and/or maintained by the surrounding tissues in vivo.
CBF1, also known as c-qin, is a member of the class 7 winged-helix transcription factor with a unique repression domain ( Kaufmann & Knochel 1996), and has been shown to exert transcriptional repressor activity ( Li et al. 1995 , 1997; for review, Vogt et al. 1997 ). Our experimental observation was consistent with this notion: The nasal CBF1 expression preluded the temporal expression of CBF2 and their expression areas did not overlap each other during development of the retina ( Fig. 1). Furthermore, misexpression of CBF1 in ovo suppressed expression of CBF2 in the temporal retina ( Fig. 4). This is consistent with the recent observation in the double temporal eyes by Müller et al. (1998) , where they showed that the expression of CBF2 in the nasally grafted temporal retina at stage 10 declined with the advance of acquisition of CBF1 during the respecification process. CBF2, in contrast, did not repress the nasal expression of CBF1 ( Fig. 4). Noteworthy is the fact that CBF1 and CBF2 are structurally classified into different subgroups of the winged-helix transcription factor ( Yuasa et al. 1996 ).
In vivo regulation of graded expression of EphA3 by transcription factors
The EphA3 is expressed in a gradient in the developing retina, and is likely to be a crucial requirement for repulsive axon guidance during the formation of retinotectal neuronal connections ( Cheng et al. 1995 ; for reviews, Drescher 1997; Flanagan & Vanderhaeghen 1998). Our initial thought was that CBF2 directly up-regulates EphA3, and CBF1 directly down-regulates EphA3. However, experimental data indicated that this was not likely to be the case.
Initial expression of EphA3 was weak and homogeneous throughout the whole optic vesicle and its graded expression was accomplished long after CBF1 and CBF2 were localized topographically ( Fig. 1). This indicates that CBF1 and CBF2 by themselves are not sufficient to confer the graded expression of EphA3 because both CBF1 and CBF2 are already expressed topographically at the stage when EphA3 begins to express non-topographically. This expression profile rather suggests that the onset of EphA3 expression itself is triggered first by an unknown transcription factor(s) in the retina to get the homogeneous expression, and subsequently the basal transcription is controlled by an additional topographic determinant(s) to attain the topographic expression at later stages during the generation of retinal ganglion cells ( Kahn 1974).
Even after completion of topographic expression, the expression domain of EphA3 in the optic vesicle was larger and more expanded to the nasal side than that of CBF2, being graded down the intensity to the nasal side ( Fig. 1). Misexpression of CBF2 influenced the expression pattern of EphA3 ( Fig. 5), although altered EphA3 expression did not always overlap with those of CBF2 misexpression. Interestingly, the modulation of EphA3 was observed when CBF2 was misexpressed at the nasal side or in the forebrain region where CBF1 is expressed ( Fig. 5), although the reason is not clear. It is likely that the CBF1-expressing nasal area may also contribute to organize the Eph gradient, reminiscent of the isthmic area at the mes-metencephalic boundary where soluble factors Wnt1 and FGF8 determine the polarity of the midbrain ( Rowitch et al. 1997 ). We postulate that CBF2 up-regulates a diffusible factor(s) that induces the expression of EphA3 topographically (see the following for further discussion).
In contrast to CBF2, CBF1 did not affect the graded expression of EphA3 in our electroporation experiments ( Fig. 5), indicating that expression of CBF1 does not directly couple with the nasal decline of EphA3. However, it is theoretically possible to speculate that CBF1 would affect EphA3 expression repressively through CBF2 because CBF1 eliminated the expression of CBF2 as mentioned above. It should be noted here that BF1 seems to be a transcription factor with dual activity: XBF-1, a Xenopus ortholog of CBF1, suppressed or induced neuronal differentiation depending on its expression dose ( Bourguignon et al. 1998 ). This might be a reason why the effect of CBF1 expression was not evident.
The homeodomain transcription factor SOHo-1 was reported to be expressed in the nasal retina ( Deitcher et al. 1994 ). We found that its topographic expression began before the topographic expression of EphA3 was established ( Fig. 2). However, the expression pattern of SOHo-1 was different from a simple gradient along the naso-temporal axis as seen for EphA3 ( Fig. 2): It appeared that the central part of the nasal region expressed less SOHo-1 than the ventral and dorsal parts. Thus, it is unlikely that this molecule directly drives the nasotemporally graded expression of EphA3. Our preliminary experiments using the in ovo electroporation technique also supported this view because SOHo-1 did not change the expression pattern of EphA3 (seven embryos; not shown), and CBF1 and CBF2 did not influence the expression pattern of SOHo-1 (two embryos each; data not shown).
Cell-autonomous and heritable expression of CBF1 and CBF2
The nasal or temporal characteristics of the immortalized retinal cells prepared from E3.5 quail embryos (corresponding to stages 23–24 in chicks) was maintained in a cell-autonomous and heritable manner ( Pollerberg & Eickholt 1995). We found that these immortalized retinal cell lines expressed not only CBF1 and CBF2 but also EphA3 according to the original character of the derived position in the retina ( Fig. 6), indicating that the expression of these genes in the respective proneural cell was imprinted before stage 23 and maintained cell-autonomously afterwards. Thus, the mitotic progenitors of ganglion cells seem to acquire positional identities prior to their cellular differentiation, and such identities are likely to be passed onto the final individual cell types. Similar to our quail cell lines, the expression of BF1 and other region- specific genes are maintained also in some immortalized cell lines derived from the embryonic mouse forebrain and midbrain ( Chun & Jaenisch 1996; Nakagawa et al. 1996 ). These results gave us the view that there might be a cellular memory system that confers fixed gene expression upon cells, once it was established ( Donoghue et al. 1992 ; Sanes 1993; Gould 1997). In the retinotectal system, such imprinting of a positional phenotype can be postulated from the results of several transplantation studies in which a cell, once determined as the nasal or temporal cell, retained its preference for the specific projection area in the tectum even in the ectopic environment ( Dütting & Thanos 1995; Dütting & Meyer 1995; Thanos et al. 1996 ).
Topographic map and EphA3
We previously demonstrated that ectopic misexpression of CBF1 and CBF2 with a retrovirus vector perturbed the topographic map in the retinotectal system along the rostrocaudal axis ( Yuasa et al. 1996 ). Similarly, EphA3 is involved in map formation because its preferred ligands ephrins A2 and A5 controlled the repulsive retinal axon behavior in the tectum ( Nakamoto et al. 1996 ; Monschau et al. 1997 ; Frisen et al. 1998 ). The CBF1, CBF2 and EphA3 were expressed as expected in the nasal and temporal cell lines derived from the respective retinal regions. Unexpectedly, however, the levels of CBF1 and CBF2 were not completely correlated with the expression of EphA3 in the retina either in normal development ( Fig. 1) or in misexpression experiments in ovo ( Fig. 5; see also Yuasa et al. 1996 ). The difference in expression pattern between CBF2 and EphA3 intimates a diffusible factor(s) as an in-between substance. The existence of such a diffusible factor(s) controlled by BF2 or controlling EphA3 has been suggested previously: In BF2-deficient mice, the BF2-deficient kidney mesenchyme failed to support epithelial morphogenesis of the ureter ducts ( Hatini et al. 1996 ; see also Bard 1996); interleukin 1β down-regulated transcription of an EphA3-related molecule in cardiomyocytes ( Li et al. 1998 ).
What would be the reason(s) for low penetrance in the effects of ectopic expression and the difference in EphA3 expression and projection phenotype? Misexpression by in ovo electroporation of the plasmid DNA is not persistent and declined after about 2 days following electroporation. In contrast, retrovirus- mediated misexpression of CBF1 and CBF2 leads to a stable mode of overexpression, even though it takes time to reach a substantial expression level, and probably could change the topographic map partly (in these experiments, only a small population of cells could attain this level before stage 11 when infection was performed at stages 9–11; see Yuasa et al. 1996 ). Another possibility is that the misexpression was still performed too late. In our experiments the timing of initiation of exogenous gene expression overlapped with the critical stage (transfection was at stage 8). The DNA transfection to earlier embryos (e.g. ≈ stage 6) might accordingly result in higher penetrance, however, efficient and specific transfection to the primordial retina is experimentally almost impossible at present.
Repulsive guidance of retinal axons via EphA3 and ephrins, which is present in the retina and tectum, respectively, may be a significant but not necessarily a predominant or the sole mechanism by which the projection map is established along the rostrocaudal axis ( McLoon 1991; von Boxberg et al. 1993 ; Savitt et al. 1995 ; Ichijo & Bonhoeffer 1998). Up until now, it is not known whether the concentration of EphA3 receptor on a single retinal axon (and its growth cone) exactly related to the target site within the topographic projection map in vivo (see Drescher 1997). Here, it should be noted that the ligands ephrins A2 and A5 were found very recently to be expressed in a high nasal-low temporal fashion also on retinal axons and their growth cones ( Marcus et al. 1996 ). Hornberger et al. (1999) showed that co-expression of ephrin-A ligands on retinal ganglion cell axons carrying EphA receptor modulated the function of these receptors. In this context, expression levels of ephrin-A ligands in the retina should also be examined next to see whether they are modified by ectopic expression of CBF1 and CBF2. Nevertheless, it seems necessary to continue the efforts to identify other topographic molecules controlled by CBF1 and CBF2, and to reveal the complex molecular network that is necessary and sufficient for projection map formation.
We thank Drs Jun-ichi Funahashi, Harukazu Nakamura and Ms Sayaka Sugiyama (Tohoku University) for technical advice concerning in ovo electroporation, and Ms Ryoko Suzuki for useful comments. We also thank Ms Akiko Kodama for secretarial assistance. This work was supported by grants from the Ministry of Education, Science, Sports, and Culture of Japan, and from CREST of the Japan Science and Technology Corporation. G. E. P. was a recipient of the Japanese-German Research Award (Japan Society for the Promotion of Science). A. M. was a fellow of the Inoue Foundation.
- 1996. A new role for the stromal cells in kidney development. Bioessays 18, 705 707.
- 1998. XBF-1, a winged helix transcription factor with dual activity, has a role in positioning neurogenesis in Xenopus competent ectoderm. Development 125, 4889 4900. , ,
- 1995. Complementary gradients in expression and binding of ELF-1 and Mek4 in development of the topographic retinotectal projection map. Cell 82, 371 381. , , ,
- 1996. Clonal cell lines produced by infection of neocortical neuroblasts using multiple oncogenes transduced by retroviruses. Mol. Cell. Neurosci. 7, 304 321. &
- 1998. Expression and tyrosine phosphorylation of Eph receptors suggest multiple mechanisms in patterning of the visual system. Dev. Biol. 193, 21 35. , ,
- 1985. The development of the retinotectal projection: An overview. In Molecular Bases of Neural Development (Eds G. M. Edelman, W. E. Gall & W. M. Cowan), pp. 389–428. John Wiley & Sons, New York. &
- 1994. Asymmetric expression of a novel homeobox gene in vertebrate sensory organs. J. Neurosci. 14, 486 498. , ,
- 1992. Mammalian muscle cells bear a cell-autonomous, heritable memory of their rostrocaudal position. Cell 69, 67 77. , , , ,
- 1997. The eph family in the patterning of neural development. Curr. Biol. 7, 799 807.
- 1995. In vitro guidance of retinal ganglion cell axons by RAGS, a 25 kDa tectal protein related to ligands for Eph receptor tyrosine kinases. Cell 82, 359 370. , , , , ,
- 1995. Transplantations of the chick eye anlage reveal an early determination of nasotemporal polarity. Int. J. Dev. Biol. 39, 921 931. &
- 1995. Early determination of nasal- temporal retinotopic specificity in the eye anlage of the chick embryo. Dev. Biol. 167, 263 281. &
- 1998. The ephrins and eph receptors in neural development. Annu. Rev. Neurosci. 21, 309 345. &
- 1998. Ephrin-A5 (AL-1/RAGS) is essential for proper retinal axon guidance and topographic mapping in the mammalian visual system. Neuron 20, 235 243. , , , , ,
- 1999. Role of Pax-5 in the regulation of a mid-hindbrain organizer’s activity. Develop. Growth Differ. 41, 59 72. , , , , ,
- 1997. Bone morphogenetic proteins (BMPs) as regulators of dorsal forebrain development. Development 124, 2203 2212. , ,
- 1997. Functions of mammalian Polycomb group and trithorax group related genes. Curr. Opin. Genet. Dev. 7, 488 494.
- 1991. Migratory paths and phenotypic choices of clonally related cells in the avian optic tectum. Neuron 6, 211 225. &
- 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49 92. &
- 1995. From tags to RAGS: Chemoaffinity finally has receptors and ligands. Neuron 15, 241 244. &
- 1996. Essential role of stromal mesenchyme in kidney morphogenesis revealed by targeted disruption of Winged Helix transcription factor BF-2. Genes Dev. 10, 1467 1478. , , , ,
- 1994. Expression of winged helix genes, BF-1 and BF-2, define adjacent domains within the developing forebrain and retina. J. Neurobiol. 25, 1293 1309. , ,
- 1993. In Molecular Probes of the Nervous System: Vol. 1. Selected Methods for Antibody and Nucleic Acid Probes. Cold Spring Harbor Laboratory Press, New York. , , , , ,
- 1999. Modulation of EphA receptor function by coexpressed ephrinA ligands on retinal ganglion cell axons. Neuron 22, 731 742. , , , , , , , , , , ,
- 1998. Differential withdrawal of retinal axons induced by a secreted factor. J. Neurosci. 18, 5008 5018. &
- 1974. An autoradiographic analysis of the time of appearance of neurons in the developing chick neural retina. Dev. Biol. 38, 30 40.
- 1994. The molecular basis of retinotectal topography. Bioessay 16, 1 11. &
- 1990. Rapid and transient decrease of N-myc expression in retinoic acid-induced differentiation of OTF9 teratocarcinoma stem cells. Mol. Cell. Biol. 10, 486 491. , ,
- 1996. Five years on the wings of fork head. Mech. Dev. 57, 3 20. &
- 1996. Expression of the Tyro4/Mek4/Cek4 gene specifically marks a subset of embryonic motor neurons and their muscle targets. Mol. Cell. Neurosci. 7, 62 74. , , , , ,
- 1995. The oncogene qin codes for a transcriptional repressor. Cancer Res. 55, 5540 5544. , , , ,
- 1998. IL-1β alters the expression of the receptor tyrosine kinase gene r-EphA3 in neonatal rat cardiomyocytes. Am. J. Physiol. 274, 331 341. , ,
- 1997. Oncogenic transformation induced by the Qin protein is correlated with transcriptional repression. Proc. Natl Acad. Sci. USA 94, 10 885 10 888. , , , ,
- 1996. Rostral optic tectum acquires caudal characteristics following ectopic engrailed expression. Curr. Biol. 6, 1006 1014. , , , , ,
- 1991. A monoclonal antibody that distinguishes between temporal and nasal retinal axons. J. Neurosci. 11, 1470 1477.
- 1996. Eph family receptors and their ligands distribute in opposing gradients in the developing mouse retina. Dev. Biol. 180, 786 789. , , , ,
- 1997. Shared and distinct functions of RAGS and ELF-1 in guiding retinal axons. EMBO J. 16, 1258 1267. , , et al.
- 1996. Manipulating gene expression with replication-competent retroviruses. Methods Cell Biol. 51, 185 218. &
- 1998. Partial respecification of nasotemporal polarity in double- temporal chick and chimeric chick-quail eyes. Mech. Dev. 74, 15 28. , , , ,
- 1997. Comparison of three non-viral transfection methods for foreign gene expression in early chicken embryos in ovo. Biochem. Biophys. Res. Commun. 230, 376 380. , , ,
- 1996. Roles of cell-autonomous mechanisms for differential expression of region-specific transcription factors in neuroepithelial cells. Development 122, 2449 2464. , , et al.
- 1996. Topographically specific effects of ELF-1 on retinal axon guidance in vitro and retinal axon mapping in vivo. Cell 86, 755 766. , , et al.
- 1995. Target preference of embryonic retina cells and retinal cell lines is cell-autonomous, position-specific, early determined and heritable. Eur. J. Neurosci. 7, 1431 1441. &
- 1995. Generation of cell lines from embryonic quail retina capable of mature neuronal differentiation. J. Neurosci. Res. 41, 427 442. , ,
- 1995. Mechanisms and molecules controlling the development of retinal maps. Perspect. Dev. Neurobiol. 3, 63 75. , ,
- 1997. Cell interactions in patterning the mammalian midbrain. Cold Spring Harbor Symp. Quant. Biol. 62, 535 544. , , , ,
- 1991. Identification of a new eph-related receptor tyrosine kinase gene from mouse and chicken that is developmentally regulated and encodes at least two forms of the receptor. New Biol. 3, 769 778. , ,
- 1993. Topographic maps and molecular gradients. Curr. Opin. Neurobiol. 3, 67 74.
- 1995. Molecular cloning of TOPAP: A topographically graded protein in the developing chick visual system. Neuron 14, 253 261. , ,
- 1997. En-2 regulates the expression of the ligands for Eph type tyrosine kinases in chick embryonic tectum. Neurosci. Res. 27, 211 217. , ,
- 1995. Longitudinal organization of the anterior neural plate and neural tube. Development 121, 3923 3933. , , , ,
- 1997. Inductive interactions direct early regionalization of the mouse forebrain. Development 124, 2709 2718. &
- 1963. Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Natl Acad. Sci. USA 50, 703 710.
- 1996. The molecular biology of axon guidance. Science 274, 1123 1133. &
- 1996. Positional determination of the naso-temporal retinal axis coincides with asymmetric expression of proteins along the anterior-posterior axis of the eye primordium. Exp. Eye Res. 63, 479 492. , , ,
- 1997. Revelations of a captive: Retroviral Qin and the oncogenicity of winged helix proteins. Virology 238, 1 7. , ,
- 1993. Guidance and topographic stabilization of nasal chick retinal axons on target-derived components in vitro. Neuron 10, 345 357. , ,
- 1992. In Situ Hybridization: A Practical Approach (Ed. D. G. Wilkinson). IRL Press, Oxford.
- 1995. Winged helix transcription factor BF-1 is essential for the development of the cerebral hemispheres. Neuron 14, 1141 1152. , , , , ,
- 1996. Visual projection map specified by topographic expression of transcription factors in the retina. Nature 382, 632 635. , , ,